PHASE II WORKING REPORT




                              by


                         WORK GROUP 2




              Atmospheric Sciences and Analysis


                      Report No. 2 - 15

                         Co-Chairmen
                  Howard L. Ferguson, Canada
                 Lester Machta, United States
                        July 10, 1981
Submitted to the Coordination Committee in Fulfillment of the
  Requirements of the Memorandum of Intent Signed by United
             States and Canada on August 5, 1980

-------
Lowell Smith, Director
Program Integration and Policy
  Staff (RD-681)
U.S. Environmental Protection Agency
Washington, D.C.  20460
Ray Robinson
Assistant Deputy Minister
Environmental Protection Service
Environment Canada
Ottawa, Ontario Canada K1A1C8
     Dear Messrs. Smith and Robinson:

          We are pleased to transmit under cover of this letter
     the PHASE II WORKING REPORT of Work Group 2 (Atmospheric
     Sciences and Analysis) as required in our terms of reference
     and work plan.  We believe that this report satisfies, in a
     scientifically responsible manner, our Phase II objectives.

          	             Sincerely,
     Lester Machta
     U.S. Chairman
     Work Group 2
         Howard E./ Ferguson
                  /
         Canadian vCJi
         Work Group 2
     cc:  R. Ewing
          E. G. Lee

-------
                            - 1 .-

                           PREFACE
     This is a "Working Report" prepared by the Atmospheric
Sciences and Analysis Work Group 2.  This group is one of
five established under the Memorandum of Intent signed by the
governments of Canada and the United States on August 5, 1980.
     This "working report" is one of a set of eleven Work
Group 2 reports in Phase 2 which represents the drawing
together of currently available information relevant to
transboundary air pollution.
     This information will be used by both governments to
develop a consensus on the nature of transboundary pollution.
     These reports contain some information and analyses that
are still preliminary in nature 7 however, they accurately
reflect the current state of knowledge as of July 10, 1981,
on the issues considered, given the resources available to
prepare these reports.  Any portion of these reports is
subject to modification and refinement as peer review,
further advances in scientific understanding, or the results
of ongoing assessment studies become available.
     More complete "final reports" dealing with a variety of
transboundary air pollution issues are expected in early
1982.  These reports will integrate the efforts of the present
"working reports" and will also incorporate editorial revisions

-------
                             -2-






                           SUMMARY






     In accordance with the Memorandum of Intent and the Work



Group 2 Work Plan, the Atmospheric Sciences and Analysis Work




Group (called Atmospheric Modeling during Phase I) continued



its charge to describe the transport and transformation of



air pollutants from their source regions to final deposition,



especially deposition in sensitive ecological regions.  During



Phase II, the Work Group activities were expanded to include



three principal areas: (1) Atmospheric Sciences Review (2)



Simulation Modeling and (3) Data Analysis Review, in order



to bring more structure and balance to the overall effort.



The results of analysis activities in the first two principal



areas are contained in separate Subgroup reports which are



summarized in Chapters 2 and 5, respectively, of this report.



The results of analysis activities in the third area are con-



tained in Chapter 3 of this report.  The Work Group and Sub-



group Reports were prepared from analyses and other material



that was reviewed extensively in monthly meetings and workshops



of Work Group 2 and its subgroups.  Three additional regional



models were added to the five Phase I models, and each of the



eight Phase II modeling groups produced a Model Profile docu-



menting the model and its results.



     The Phase II Atmospheric Sciences Review addressed four



critical issues in long-range transport modeling, namely: (1)



a state-of-the-art review of our understanding of sulfur and

-------
                             -3-

nitrogen chemistry in the atmosphere; (2) the evidence for
trends in precipitation composition and deposition; (3) the
seasonal dependence of deposition and chemical transformation
rates for sulfur and nitrogen compounds and the adequacy of
current parameterization methods used in regional models for
these processes; and (4) the global distribution of acidic
precipitation and its implications for Eastern North America.
     The review of sulfur chemistry concluded that the rate
of homogeneous gas phase conversion of S02 to 864 is dominated
by free radical reaction processes and the concentration of
the important free radicals is dependent upon many factors,
especially the concentration of volatile organic compounds
and nitrogen oxides, the temperature and the solar radiation
intensity.  Our knowledge of heterogeneous oxidation of S02
is less complete, but indicates that liquid phase catalyzed
oxidation by the manganese ion, carbon, and hydrogen peroxide
all could be potentially important.  However, there is uncer-
tainty about the actual availability of these catalyzing
substances in ambient fine particulate matter.
     The review of nitrogen chemistry concluded that the fate
of nitric acid in the atmosphere is not well understood, but
it would still be useful to apply results of our limited
understanding of nitrogen chemistry to exploratory modeling
exercises.  Consequently some preliminary efforts have been
made to treat nitrogen reactions in a relatively crude manner

-------
                             -4-

in existing transport models.  This preliminary experience
with modeling nitrogen concentrations and deposition on a
regional scale leads to the conclusion that greater temporal
and spatial resolution for nitrogen modeling is necessary
and more extensive and reliable monitoring information is
required for validation purposes.
     The review of the evidence for trends in precipitation
composition and deposition concludes that, in spite of the
difficulties with the data base and the controversy on the
subject, the data do suggest expansion of the region covered
by acidic rainfall, especially into the southeastern and mid-
western portions of the U. S. and the southeastern portions
of Canada.  In this regard, the Modeling Subgroup decided to
focus its Phase II efforts on the present situation and not
to apply the models to historic data because of the uncertain-
ties in past monitoring measurements and other necessary
input data, particularly the historic emissions fields.
     The review of the seasonal variation of deposition and
chemical transformation rates concluded that many of the
parameters in regional models could be strongly dependent
upon latitude during winter months and recommended that not
only seasonal variability, but also the spatial variability
be taken into account.  This review provided some specific
suggestions for the participating modelers for Phase III.

-------
                             -5--

     Th e review of the global distribution of pH and its
implications concluded that all precipitation appears to
contain both acidic and basic materials in small quantities.
Thus, a given pH value can be associated with various combi-
nations of acidic and alkaline constituents in the precipita-
tion, depending upon the region of the hemisphere under
consideration.  There is evidence that the "hemispheric
background" value of precipitation pH (i.e., the average
annual value at remote sites farthest removed from the three
major anthropogenic source regions in the northern hemisphere)
is significantly lower than the idealized "clean atmosphere"
value of 5.6.  This average hemispheric precipitation pH
value may now be closer to 5.0 than to 5.6.
     However, a significant region of Eastern North America
is experiencing an average annual precipitation pH of less
than 4.2, that is, about 10 times as acidic as this estimated
hemispheric background value.  The total deposition of acid
(H+ ions) and sulfate (sum over all wet precipitation
events and dry deposition processes) needs to be considered
in assessing effects on sensitive ecosystems.  More analysis
of data are required to determine what proportion of observed
"background pH levels" is due to natural sources or to the
residual effect of man-made sources far upwind.

-------
                             -6--





     Th e Phase II Data Analysis review covered annual, sea-



sonal, and episode deposition monitoring results, as well as



their interpretation in conjunction with emission inventories,



ambient concentration data, and trajectory calculations.



This review found the highest precipitation acidity on an



annual basis in the northern hemisphere over (1) eastern North



America, (2) western Europe and (3) Japan.  Near neutral pre-



cipitation, frequently in excess of pH 6, is found over the



large continental areas of western North America and Asia.



The cause of the slightly acidic precipitation along the west



coast of North America has not yet been completely explained,



but may be due to either anthropogenic sources or the release



of biologically-produced organic sulfur compounds from the



Pacific Ocean surface, or both.



     The zone of maximum acidity in Eastern North America



stretches in a corridor through Ohio and Pennsylvania into



Southern Ontario.  Available concentration data at locations



in eastern North America well removed from major source regions



generally indicate a summer sulfate maximum and a winter S02



maximum with highly episodic behavior of both on a daily basis.



In addition, calculated dry depositions of sulfur are found to



be of comparable magnitude to wet sulfur depositions especially



close to source regions and in the winter season.

-------
                             -7-

     Recent interpretations of both concentration and deposi-
tion monitoring data using trajectory calculations indicate
that maritime tropical air masses from the U.S. are the
principal conveyors of elevated concentrations and depositions
to the extreme northeastern U.S. and southeastern Canada, as
opposed to continental polar air masses from Canada.  It was
recognized in making these interpretations that source-receptor
relationships, based upon calculations of transport and chemical
transformations between probable sources and the receptor of
interest and upon event data at single monitoring stations,
are not always straightforward and are subject to uncertainty.
Little can be done in the immediate future to resolve some
of these uncertainties, particularly those associated with.
trajectories in some precipitation systems and the variable
rate of chemical transformations from event to event; while
others can be reduced by analysis of data at more sites and
for longer periods of record.
     Finally, the data analysis review concluded that within
eastern North America, natural sources of sulfur within the
region are unimportant compared to anthropogenic sources.
Somewhat more significant are the background sulfur concen-
trations that are transported to eastern North America from
the Pacific and Carribean Oceans, the Atlantic Ocean south
of 30°N, and the arctic region.  These manifestations of the
hemispheric background contribution to acid deposition are

-------
                             -8-

still considered to be small in comparison with the local
and long-range transport impacts in eastern North America.
The sulfur is scavenged by orographic precipitation from
Pacific air masses as they cross the coastal mountains leaving
concentrations close to the "hemispheric background levels"/
whereas the Carribean and artic air masses may contain some-
what elevated sulfur concentrations relative to those in
the modified Pacific air masses.  The predominant sources of
elevated sulfur concentrations in arctic air masses in the
winter are thought to be, in order of decreasing importance,
Siberia, Europe, and Eastern North America.
     The Phase II modeling included (1) a further analysis of
regional air quality modeling in general and the role of
modeling in the development of an air quality agreement,  (2)
an intercomparison and evaluation of models using standardized
inputs and evaluation criteria, (3) an analysis of seasonal
variabilities and uncertainties in the transfer matrices for
sulfur oxides; and (4) generation of preliminary transfer
matrices for primary sulfate and nitrogen oxides emissions.
The analysis and its role in appropriate applications of
modeling indicated that modeling predictions are expected to
deviate to some degree from actual monitoring measurements.
For practical reasons, models do not incorporate all of our
current understanding of the relevant physical processes,
which itself is incomplete.  Furthermore, our available

-------
                             -9-

monitoring data bases are insufficient to compute the ensemble
average which the model is designed to predict.  The uncer-
tainties in model predictions may be quantified from the
differences between model predictions and observations.
     Although the application of regional models is constrained
by these uncertainties inherent in their calculations, such
constraints can be alleviated to a significant degree by
requiring the modelers to quantify the relevant uncertainties,
and by taking these into account in any application.  Some of
the uncertainties in the transfer matrix elements can be
assessed by analyzing the transfer matrices from more than
one model and by using probabilistic techniques of analysis
which will be developed during Phase III.  Other uncertainties
may be quantified after further model evaluation efforts are
completed.
     While there is still no general agreement within the
modeling community as to the proper method and the statistics
for intercomparison and evaluation of models, the Modeling
Subgroup has made significant strides in selecting a common
basis for performing these tasks for the eight participating
modelers.  In Phase II, a complete set of evaluation statistics
was computed by only one modeler, while the monthly and annual
residuals were computed at 9 to 20 sites by three of the eight
modelers.  The other four modelers are expected to complete this
preliminary evaluation by September 1981.  No single model has
emerged as clearly superior or inferior to the others from this

-------
                             -10-

first of three rounds of model evaluation.  The evaluation
has primarly served to reveal (1) the deficiencies in the
monitoring data bases, (2) the need for some changes in input
parameters for some of the models/ and (3) the need to use at
least one more year of independent data for model evaluation.
     The seasonal transfer matrices computed with several
of the models were too preliminary to draw any general con-
clusions, but indicate that seasonal variability can be
significant and should be investigated further during Phase
III.  In addition most of the models completed detailed
sensitivity analyses by varying each input parameter sepa-
rately within the limits normally used for long-range trans-
port modeling with either (1) actual meteorology and emissions
or (2) a hypothetical source-receptor situation and simulated
meteorology.  These sensitivity analyses, which are documented
in the individual Model Profiles, provide a more complete
understanding of the workings of each model and are useful
for incorporating uncertainty into the analysis process.
     The annual transfer matrices for Phases I and II were
intercompared and, interestingly, the use of standardized
inputs did not reduce the range of variation among models in
some of the transfer matrix elements expressed in two of the
three standard forms, namely absolute values or normalized
by unit emissions of sulfur.  However, the transfer matrix
elements expressed as percentage contribution from a source

-------
                             -11-

region to a receptor area were generally in much better agree-
ment among the models.  Since additional refinements will be
made to most of the models and a new set of source-receptor
regions will be used during Phase III, it is premature to
draw any general conclusions at this time.
     These variations in transfer matrix coefficients reflect
the current uncertainties in how best to parameterize all
the physical processes and are the result of different
approaches by independent modelers.  Although the desirability
of a single/ unified transfer matrix was recognized/ the
Modeling Subgroup has reservations about the generation and
application of a unified transfer matrix at this time because
no matter how it is generated its interpretation is subject
to some question.
     Since the modeling of nitrogen oxides chemistry is in
its infancy and because there is so little data available
from which to select the model parameterizations, any modeling
at this stage cannot yield more than some educated "first
estimates" of the long-range transport of nitrogen oxides.
The preliminary results from three models, in terms of transfer
matrices and comparison to monitoring data/ indicate the primi-
tive nature of current efforts.  Therefore/ at this stage,
these should not be used for analysis efforts.
     A preliminary transfer matrix for primary sulfate emis-
sions was generated during Phase II to assess the relative
contribution to acid deposition of primary sulfate emissions

-------
                             -12-





from oil-fired and coal-fired combustion sources compared to



that from secondarily formed sulfates.  Newly published pri-



mary sulfate emission factors from large package boilers were



utilized by one regional transport model to develop this



comparative analysis.  The analysis indicated that primary



sulfate concentrations exceeded secondary sulfate concentra-



tions in the New York City area in the winter season, while



secondary sulfate concentrations exceeded primary sulfate



concentrations in all other areas in the winter and in all



areas during the summer.   Additional modeling of the contri-



bution of primary sulfate emissions will be performed during



Phase III using an improved emissions inventory and an attempt



will be made to evaluate the results against monitoring



data.



     The principal objective for Phase III will be to refine



and consolidate the. information provided in the Phases I and



II reports.  In addition, some attention will be directed to



other (non-acid deposition) transboundary air pollution



issues that are likely to be considered in transboundary air



quality negotiations.  The Phase III work will be performed



by four subgroups, with a Local Source Analysis Subgroup to



be added to the three existing subgroups.  The Work Group



will also finalize its proposals for the research, modeling



and monitoring element of the agreement which appear in a



preliminary form in this report.

-------
                                        -13-
                                LIST OF CONTRIBUTORS


                 This report was prepared  by members of Work Group 2 as

            listed below.  The  first two authors  listed carried the

            primary responsibility  for writing  the  chapters.   Reviewers

            provided comments on final draft sections.   In all cases

            Canadian and U.S. Work  Group members  worked closely on the

            preparation of individual chapters  and  on the  final integra-

            tion of the complete report.   Drs.  Brand L. Niemann and James

            W.S. Young were responsible for  coordinating the  preparation

            of the report.
Chapter

 1


 2
    Title
Introduction
1. L. F. Smith
2. H. L. Ferguson
Atmospheric     1. J. Miller
Science Issues  2. P.W. Summers
in LRT Modeling

Monitoring Results 1. J. Miller
and Interpretation 2. J. W. S. Young
Author(s)

   3. B. Niemann
   4. J. W. S. Young

   3. H. L. Ferguson
   4. P. Altshuller
                                              3. B. Niemann
Reviewer(s)

   H. Martin
   J. Blanchard

   L. Machta
   R. Ball
                                          G. Paulin
                                          F. Burmann
                                          N. laulainen
        The Role of
        Models in the
        Development of
        Omission Control
        Strategies and
        An Air Quality
        Agreement

        Summary of
        Selected Models
        and Their Inter-
        comparison/
        Evaluation
                   1. R. Ball
                   2. P. K. Misra
                     3.  H. L. Ferguson   G. Van Volkenburg
                                          R. Morris
                   1. J. W. S. Young   3.  M. P. Olson
                   2. B. Niemann       4.  T. Clark
                                           P. K. Misra
                                           K. Demerjian.

-------
                                          -14-
Chapter     Title

 6      Phase II Improved
        Emission Inventory

 7      Source-Receptor
        Relationships for
        Sulfur Oxides

 8      Analyses of Transfer
        Matrices for Sulfur
        Oxides

 9      Preliminary
        Source-Relation-
        ships for Nitrogen
        Oxides

10      Preliminary
        Source-Receptor
        Relationships for
        Primary Sulfates
   Author(s)

1. F. Vena
2. B. Niemann

1. T. Clark
2. M. Olson
             Reviewer(s)

                  G. Paulin
                  L. Smith

3. P. Altshuller  B. Power
                  R. Ball
1. D.. Ball       3. B. Niemann
2. M. P.. Olson
1. M. P. Olson
2. K.. Demerjian
1. G. Van Volkenburg 3. F. Vena
2. B. Niemann
                  R. Harrington
11      Conclusions,
        Recommendations,
        and Phase III
        Work
1. J. Miller
2. G. Van Volkenburg
3. H. L. Ferguson
4. L. Machta
12      Preliminary
        Proposals for
        Research, Model-
        ing, and Moni-
        toring Element
        of the Agreement
1. P. W. Summers
2. L. Smith
     3. J. W. S. Young  L. Machta
     4. B. Niemann      H. L. Ferguson

-------
                             -15-


                       ACKNOWLEDGEMENTS


     Work Group 2 wishes to acknowledge the substantial

contributions made by the participating modelers, non-government
                i
scientists and support staff to the completion of this report.

     The names of these individuals and their principal

contributions are listed below.


 Name                        Principal  Contribution

Man's Lusis                   Chapter 2
Lou Shenfeld                  Chapters  2 and 7
Don Lewis                     Chapters  4 and 8
Barbara Ley                   Chapters  3, 4 and 8 and
                               OME Model Profile
Ed Pechan                     Chapters  6 and 10 and
                               General  Computer Support
A. Veflkatram                  Chapter 4
C. Bhumralkar                 Chapter 8
Jan Bottenheim                Chapter 9
Jack Shannon                  Chapter 10 and
                               ASTRAP Model Profile
Len Barrie                    Chapter 3
Kurt Anlauf                   Chapter 3
Eva Voldner                   Chapters  5 and 8 and AES-LRT.Mode 1
                               Profile
RodShaw                      Reviewofentirereport
Carolyn Acklin and            Word Processing
Jacque'Hawkins

-------
                                 - 16 -


                           TABLE OF CONTENTS

                                                                 PAGE

"' PREFACE	       1

 SUMMARY	       2

 LIST  OF  CONTRIBUTORS	      13

 ACKNOWLEDGEMENTS	      15

 LIST  OF  FIGURES	      20

 LIST  OF  TABLES	      24

 LIST  OF  ACRONYMS	      27

 1.  INTRODUCTION	      1-1

    1.1   General	      1-1
    1.2   Phase  II  Activities	      1-1
    1.3   Phase  II  Report	      1-4
    1.4   Phase  III Activities	      1-6

 2.  ATMOSPHERIC SCIENCE  ISSUES  IN LRT MODELING	      2-1

    2.1   Introduction	      2-1
    2.2   Sulfur and Nitrogen Chemistry in'
          LRT  Models	      2-1
    2.3   Trends in Precipitation Composition
           and Deposition	      2-3
    2.4   Seasonal  Dependence of Atmospheric Deposition
           and Chemical Transformation Rates for
           Sulfur and Nitrogen Compounds	      2-5
    2.5   The  Global Distribution of Acidic Precipitation
           and Its  Implications  for Eastern
           North America	      2-9

 3.  MONITORING  RESULTS AND INTERPRETATION		      3-1

    3.1   North  American Precipitation
          Acidity	      3-1
    3.2   Temporal  Variations	      3-3
    3.3   Classification According to Air
           Parcel Origin	      3-13
    3.4   Episodes  of Long-Range Transport...	      3-33
    3.5   Background Sulfur in North America
           East  of  the Rocky Mountains	      3-40
    3.6   Conclusions	      3-51

-------
                                 - 17 -
                                                                PAGE

4.   THE ROLE OF MODELING IN THE DEVELOPMENT OF
      EMISSION CONTROL STRATEGIES AND AN AIR
      QUALITY AGREEMENT	      4-1

    4.1  Introduction	      4-1
    4.2  Uncertainties in Model Predictions	      4-5
    4.3  Use of Model Results in Decision Making	      4-6
    4.4  Transfer Matrices	      4-9

5.   SUMMARY OF SELECTED MODELS AND THEIR INTER-
      COMPARISON/EVALUATION	      5-1

    5.1  Summary of Model Profiles	      5-1

         5.1.1  AES-LRT	      5.1-1
         5.1.2  ASTRAP	      5.2-3
         5.1.3  ENAMAP	      5.3-8
         5.1.4  OME-LRT	      5.4-11
         5.1,5  RCDM-2	      5.5-15
         5.1.6  CAPITA Monte Carlo	      5.6-18
         5.1.7  MEP-TRANS	      5.7-22
         5.1.8  UMACID	      5.8-29

    5.2  Intercomparison and Evaluation	      5.2-1

6.   PHASE II IMPROVED EMISSION INVENTORY	      6-1

    6.1  Emissions	      6-1
    6.2  Source-Receptor Areas	      6-5

7.   SOURCE-RECEPTOR RELATIONSHIPS FOR SULFUR OXIDES	      7-1

    7.1  Introduction	      7-1
    7.2  Source-Receptor Matrix Description	      7-1
    7.3  Intercomparison of Phase I and Phase II
          Transfer Matrices	      7-8
    7.4  Comparison of Model Estimates with
          Observations	      7-12

8.   ANALYSES OF TRANSFER MATRICES FOR SULFUR OXIDES	      8-1

    8.1  Introduction	      8-1
    8.2  Seasonal Variations	      8-1

         8.2.1  AES-LRT Results	      8-1
         8.2.2  ASTRAP Results	      8-7
         8.2.3  Summary	      8-10

-------
                                 -  18 -
                                                                 PAGE
    8.3  Model Sensitivity to Parameters	       8-11

         8.3.1  AES-LRT Results	       8-12
         8.3.2  OME-LRT Results	       8-15
         8.3.3  MEP-TRANS Results	       8-20
         8.3.4  Other Model Results	   '    8-24

9.  PRELIMINARY SOURCE-RECEPTOR RELATIONSHIPS FOR
      NITROGEN OXIDES	       9-1

    9.1  Introduction	       9-1
    9.2  Parameterization	       9-4
    9. 3  Modeling Results	       9-10
    9.4  Conclusions and Recommendations	       9-14

10. PRELIMINARY SOURCE-RECEPTOR RELATIONSHIPS
      FOR PRIMARY SULFATES	      10-1
                                              /
    10.1 Introduction	;	      10-1
    10.2 Emissions	      10-3
   " 10. 3 ASTRAP Model Results	      10-5

11. CONCLUSIONS, RECOMMENDATIONS, AND PHASE III
      WORK PLAN	      11-1

    11.1 Conclusions and Recommendations	      11-1
    11.2 Phase III Work Plan	      11-9

12. PRELIMINARY PROPOSALS FOR RESEARCH, MODELING, AND
      MONITORING ELEMENT OF THE AGREEMENT	      12-1

    12.1 Introduction	      12-1
    12.2 Atmospheric Processes Research	      12-2
    12.3 Monitoring	      12-3
    12.4 Modeling	      12-4
    12.5 Atmospheric Science Assessment	      12-5

REFERENCES	       R-l

ADDENDUM A -  MODEL CRITIQUE BY U.S. CHAIRMAN OF
                WORK GROUP 2 LESTER MACHTA	       A-l

ADDENDUM B -  RESPONSE TO ADDENDUM A BY CANADIAN
                CHAIRMAN OF WORK GROUP 2
                HOWARD FERGUSON	       B-1

-------
                                 - 19 -
                                                                PAGE

APPENDICES	

1 - Work Group 2 Terms of Reference and
    Additional Guidance	      A.1-1
2 - Membership of Work Group 2	      A. 2-1
3 - Glossary of Terms	      A.3-1
4 - Compilation of Attendance/ Agenda and Minutes
      for Modeling Subgroup Workshops and
      Work Group 2 Meetings	      A.4-1
5 - List of Work Group 2 Reports and Other Documents	      A.5-1
6 - A Simple Example of the Role of Models in the
      Development of Emission Control Strategies	      A.6-1

-------
Figure 3-1,
Figure 3-2.


Figure 3-3.



Figure 3-4.



Figure 3-5.
Figure 3-6.



Figure 3-7.


Figure 3-8.


Figure 3-9.




Figure 3-10.
                            - 20 -

                       LIST OF FIGURES
Annual Average pH of Precipitation in North
America during 1979 Based on Observations by
Canadian APN and CANSAP Networks and American
MAP3S and NADP Networks.

Estimated Distribution of the pH For Rain Water
in the Northern Hemisphere.

Sites in the Canadian Air and Precipitation
Monitoring Network (APN) At Which Daily Air
and Precipitation Samples are Collected.

Temporal Variations of the Monthly Average Con-
centration of Atmospheric Sulfate and Sulfur
Dioxide At APN Sites During 1979.

Top:  The Fraction of Total Airborne Sulfur
Existing As Sulfur Dioxide At Long Point On the
North Shore of Lake Erie on a Daily Basis
(19-78-1979).  Bottom:  The Concentration of
Total Airborne Sulfur At Long Point on a Daily
Basis.

The Temporal Variation of Monthly Precipitation-
Amount Weighted Mean Concentrations of Four Major
Ions at APN Sites During 1979.

Monthly Mean Deposition - Weighted Concentrations
for H+, S04= and N03~ at the MAP3S Stations.

The Temporal Variation of Monthly Wet and Dry
Deposition of Oxides of Sulfur at APN Sites.

Trajectory Wind Roses At the APN Sites For Par-
ticulate Sulfate Based On Two Day Back
Trajectories At 925 mb(about 800 meters) During
January - December 1979.

Trajectory Wind Roses at APN Sites For .Sulfur
Dioxide Based On Two Day Back Trajectories At
925 mb (about 800 meters) During January -
December 1979.
Page

3-2
3-4


3-5



3-6



3-9
3-11



3-12


3-14


3-18
3-19

-------
                             - 21 -
Figure 3-11.
Figure 3-12.
Figure 3-13.
Figure 3-14.
Figure 3-15.
Figure 3-16.
Figure 3-17.
Figure 3-18.
Figure 3-19:
Trajectory Wind Roses At APN Sites For Rain
pH Based On Two Day Back Trajectories At 925
mb(about 800 meters) During January - December
1979.

Trajectory Wind Roses At APN Sites For Rain
Sulfate Based on Two Day Back Trajectories at
925 mb(about 800 meters) During January -
December 1979.

Trajectory Wind Roses at APN Sites for Rain
Nitrate Based on Two Day Back Trajectories at
925mb (about 800 meters) During January -
December 1979.

Top:  Precipitation Volume, and Bottom:
Hydrogen Ion Total Wet Depostion, At Whiteface
Mountain, New York, Based On Two Day Back
Trajectories In the Boundary Layer Assigned To
30° Sectors.

Top:  Sulfate, and Bottom:  Nitrate Total Wet
Deposition, At Whiteface Mountain, New York,
Based On Two Day Back Trajectories In the
Boundary Layer Assigned To 30° Sectors.

Top: Calcium, and Bottom:  Ammonium Total Wet
Depostion, At WhiteFace Mountain, New York,
Based On Two Day Back Trajectories In the
Boundary Layer Assigned To 30° Sectors.

Top:  Precipitation Volume, and Bottom:  Hydrogen
Ion Total Wet  Deposition at Whiteface Mountain,
New York, Based On Two Day Back Trajectories
In the Boundary Layer Assigned To 30° Sectors.

Top:  Sulfate, and Bottom:  Nitrate Total Wet
Deposition, At.Whiteface Mountain, New York,
Based On Two Day Back Trajectories In the
Boundary Layer Assigned To 30° Sectors.

Surface Weather Map Showing Large High Pressure
Area Centered Over the Midwestern States On
Feburary 19, 1979.
3-21
3-22
3-24
3-25
3-26
3-28
3-29
3-35

-------
                           - 22 -
Figure 3-20:



Figure 3-21.



Figure 3-22.


Figure 3-23.



Figure 3-24.



Figure 3-25.



Figure 3-26.




Figure 3-27.



Figure 3-28.




Figure 3-29.



Figure 6-1.



Figure 7-1.



Figure 8-1.
                                                   Page

Surface Weather Map Showing the Movement of the    3-36
Large High Pressure Area From the Midwest To
the East Coast On Feburary 20, 1979.

Top:  Daily Sulfate, and Bottom:  Sulfur Dioxide   3-37
Concentration (ug m~3) At the Experimental Lakes
Site.

Back Trajectories at 1000mb During February        3-38
16-20, 1979.

pH of the First Rain Sample At Each Station In     3-41
the OSCAR Program Network During the Storm Event
of April 22-24,  1981.

Change of pH With Time At Four Selected Stations   3-42
In the OSCAR Program Network During the Storm
Event of April 22-24, 1981.

Locations of Sites In the Arctic Aerosol Sampling  3-48
Network (M- Mould Bay, I- Igloolik and A- Alert)
and of the American Sampling Site (B- Barrow).

Weekly-Average Excess-Sulfate (non-sea salt)       3-49
Concentrations in the Atmosphere At Mould Bay
and Igloolik As  Well As of Monthly Average Cloud
Cover in the Arctic.

Diagram of Weekly Average and Minimum S02 and      3-50
804" Concentrations At the Experimental Lakes
Site Between November 1978 and December 1979.

Acidity of Fresh Snowfall At Mould Bay Estimated   3-52
From Measured Sulfate Concentrations, Hydrogen
Ion to Sulfate Ratios in Aerosols and a
Scavenging Ratio of 2 x 10^ (by volume).

Seasonal Fluctuations In the pH of the Snow In     3-53
the Agassiz Ice Cap Ellesmere Island For Each
Year Marked.

Canadian Province and Sub-Province Regions and     6-7
U. S. State and  Multi-State Regions for the
Phase II Transfer Matrices.

Map of Eastern North America Showing the 11        7-2
Major Source Regions and 9 Sensitive Areas
Used in the Phase II Transfer Matrices.

Idealized Source - Receptor Geometry Used for      8-16
OME-LRT Model Sensitivity Studies.

-------
                             - 23 -
Figure 9-1.

Figure 9-2.


Figure 10-1.
Figure 10-2.
Figure 10-3.
              Gas Phase Nitrogen Oxides Chemistry.

              The Kinetic Scheme for NOX Modeling Used  by  the
              CAPITA Model.

              Isopleths of Summer  (July-August) Primary
              Sulfate Concentrations  (jug m~3) Simulated by
              the ASTRAP Model  (Maximum Concentration  3.7
              Jig m"3).

              Isopleths of Summer  (July-August) Secondary
              Sulfate Concentrations  (jug m~3) Simulated by
              the ASTRAP Model  (Maximum Concentration  14.9
                 m~3).
Isopleths of Winter  (January-February)
Secondary Sulfate Concentrations  (
-------
                            - 24 -
                         LIST OF TABLES
                                                   Page
Table 3-1.



Table 3-2.



Table 3-3.


Table 3-4.


Table 3-5.


Table 6-1.



Table 6-2.



Table 6-3.


Table 6-4.


Table 6-5.


Table 6-6.

Table 7-1.

Table 7-2.
Estimated Monthly Average Dry Depostion            3-15
Velocities (cm s"1) of Sulfur Dioxide at APN
Sites.

Estimated Monthly Average Dry Depostion            3-15
Velocities (cm s"1) of Particulate Sulfate
at APN Sites.

Definition of Concentration and pH Ranges for      3-17
the Trajectory Cases.

Major Source-Receptor Analyses Using Trajectories  3-32
or Storm Tracks.

A Breakdown of Natural and Anthropogenic Sulfur    3-44
Emissions (Tg.S. yr"1)

Phase II Improved United States and Canadian       6-2
SC>2 Emissions on a State and Province Basis
(Kilotonnes/Yr.) - 1980.

Phase II United States and Canadian SC>2   "        6-3
Emissions for the 63 ARMS Areas (Kilotonnes/
Yr.) - 1980.

Combined U. S. - Canadian Top 50 Sources of SC-2    6-4
Emissions - 1980.

Phase II Targeted Sensitive Areas for Work Group   6-6
2 Modeling.

Canadian Regions for Phase III Transfer            6-8
Matrices.

U. S. Regions for Phase III Transfer Matrices.     6-9

Key to 11 Source Regions and 9 Sensitive Areas.    7-3

Phase I Transfer Matrix of Annual Wet Deposition   7-5
of Sulfur (kg.S. ha."1 yr.""1) Per Unit Emission
(Tg.S.yr.-1)

-------
                           - 25 -
Table 7-3.



Table 7-4.



Table 7-5.


Table 7-6.



Table 8-1.


Table 8-2.


Table 8-3.



Table 8-4.


Table 8-5.



Table 8-6.
Table 8-7.



Table 8-8.


Table 8-9.
                                                   Page

Phase I Transfer Matrix of Annual Wet Deposition   7-9
of Sulfur fkg.S. ha."1 yr."1) Per Unit Emission
(Tg.S.yr.'1)

Selected Transfer Matrix Elements Values of        7-11
Annual Wet Sulfur Deposition from Phase I and
Phase II.

Significant Changes in Phase I and II RCDM and     7-13
ASTRAP Annual Wet Sulfur Depositions.

Model Estimates and Observations of Annual Wet     7-14
Sulfur Deposition (kg.S.ha."1 yr."1) at the Nine
Targeted Sensitive Areas.

AES-LRT Model Parameters Used in the Sensitivy     8-3
Study.

Seasonal Variations In The AES-LRT Model Transfer  8-4
Matrices of Absolute Values (1978).

Seasonal Variations In The AES-LRT Model Transfer  8-6
Matrices of Per Cent of Total or of Annual
Average  (1978).

Seasonal Variation in Parameter Values For the     8-8
ASTRAP Model.

January and July 1978 ASTRAP Model Estimates of    8-9
SC>2 Concentrations and Wet Sulfur Depositions
at the Nine Targeted Sensitive Receptors.

Sensitivity Index - Fractional Change in Wet       8-13
Sulfur Deposition As a Function of Fractional
Change in Parameter Value Annual (din Dep/dln
Parameter).

OME-LRT Model Sensitivity of the Wet Sulfur        8-17
Deposition Factor for the Idealized Source-
Receptor Geometry Shown in Figure 8-1.

Range of Parameter Variation for the MEP-TRANS     8-21
Model Sensitivity Analysis.

Sensitivity of Wet 804  Deposition to Variations   8-23
in MEP-TRANS Model Parameters.

-------
                             - 26 -
Table 9-1.


Table 9-2.


Table 9-3.


Table 9-4.



Table 9-5.


Table 10-1.


Table 10- 2.
Table 10-3.
                                                   Page

NO  Kinetic Rate Constants  (h"1) Used in the       9-1
CAPITA Model.

Parameter Choice For 1978 MEP-TRANS Model          9-2
Simulations of NOX.

Deposition Parameters for NOX Chemistry Used in    9-3
the AES-LRT Model.

Normalized Transfer Matrix  for Total Nitrogen      9-12
from the MEP-TRANS Model (kg.N. ha.-iyr."1 per
Tg. N. yr."1)

Correlations Between Observed and Predicted        9-13
N03~ From the AES-LRT Model.

Sulfate Emission Factors and Sulfur Oxide         10-5
Emission Rates.

Transfer Matrix for January 1978 Sulfur Dioxide   10-14
and Sulfate Concentrations  (jug m~^) from the
ASTRAP Model Using the Phase III State/Province
S02 Emission Inventory and  Primary Sulfate
Emission Factors in Table 10-1.

Transfer Matrix for July 1978 Sulfur Dioxide and  10-16
Sulfate Concentrations (/ag m"~3) from the ASTRAP
Model Using the Phase III State/Province S02
Emission Inventory and Primary Sulfate Emission
Factors in Table 10-1.

-------
                              -27-


                       LIST OF ACRONYMS


BRCG or RCG - Bilateral Research Consultation Group (U.S.-Canada)

MAP3S/RAINE - Multi-State Atmospheric Power Production Pollution
  Study/Regional Acidity of Industrial Emissions

EPA - Environmental Protection Agency (U.S.)

AMS - American Meteorological Society

LRTAP - Long-Range Transport of Air Pollutants

AES - Atmospheric Environment Service (Canada)

CMC - Canadian Meteorological Centre

SURE - Sulfate Regional Experiment (EPRI)

UTM - Universal Transverse Mercator

EPRI - Electric Power Research Institute

ENAMAP - Eastern North American Model of Air Pollution

ASTRAP - Advanced Statistical Trajectory Regional Air Pollution
  Model

1-D - one-dimensional

EURMAP - European Regional Model of Air Pollution

SRI — formerly Stanford Research Institute, now SRI International, Inc

UNIVAC - name of a computer company

NEDS - National Emissions Data System (EPA)

OME - Ontario Ministry of the Environment

GCA - formerly Geophysics Corporation of America now GCA Corporation

NOAA/ATDL - National Oceanic and Atmospheric Administration/
  Atmospheric Transport and Dispersion Laboratory

RCDM - Regional Climatological Dispersion Model

TRI - Teknekron Research, Inc.

-------
                             -28-





CAPITA - Center for Air Pollution Impact and Trend Analysis



MEP - Meteorological and Environmental Planning, Ltd.



TRANS - Transport of Regional Atmospheric Nitrogen and Sulfur



APN - Air and Precipitation Monitoring Network  (Canada)



CANSAP - Canadian Network for Sampling Precipitation



NADP - National Atmospheric Deposition Program  (U.S.)



ACID - Atmospheric Contribution to Inter-Regional Deposition



DOE - Department of Energy



ARMS - Acid Rain Mitigation Studies



RMS  - Root-mean-square-error



RMSB - Standard deviation of residuals

-------
                          Chapter 1

                         INTRODUCTION
1.1  General
     The Atmospheric Sciences and Analysis Work Group was

established under the Memorandum of Intent in order to provide

information, based on cooperative atmospheric modeling and

analysis of monitoring network and other data/ which would

lead to a further understanding of the transport of air pollu-

tants between source regions and sensitive areas.  In addition,

the Group was to prepare proposals for the "Research, Modeling

and Monitoring" element of an agreement.  The Terms of Reference

of the Group, Work Group Membership, and Glossary of Terms are

contained in Appendices 1, 2 and 3, respectively.

1.2  Phase II Activities

     During Phase II the Work Group activities have been

structured in three activity areas with purposes as follows:
     1.  Atmospheric Sciences Review - assess the appropriate-
         ness of the methods and assumptions used in regional
         models to quantify source-receptor relationships;

     2.  Simulation Modeling - document, evaluate, inter-
         compare and apply available practical regional
         models; and

     3.  Data Analysis Review - use data to establish indepen-
         dentlyfT)the usefulness of regional models and
         (2) the validity of computed source-receptor
         relationships.

Appendix 4 contains a listing of the agendas, attendance rosters,

and minutes of the several Modeling Subgroup workshops and

Work Group 2 meetings held during Phase II, while Appendix 5

contains a listing of Work Group 2 Phase I and II reports.

-------
                             1-2.





     Although many substances may undergo transboundary



atmospheric transport and have harmful effects upon either



the atmosphere or surface receptors, acid deposition has been



the phenomenon of primary concern for the first two phases



of our Work Group activities.  As a consequence, the highest



priority has been given to the study of oxides of sulfur and



nitrogen, the main precursors of acidic deposition.  Because



of the analysis needs of the other Work Groups to develop con-



trol measures that would be effective in reducing transboundary



acid deposition effects, significant emphasis has been



placed on the development of the "transfer matrix" concept.



The transfer matrix specifies the contribution of individual



source areas to receptor areas of interest.  It assumes a



linear relationship between sources and receptors, and is



generated from mathematical models of long range transport



of air pollutants.  Transfer matrices have only been con-



structed, to date, for sulfur compounds for seasonal and



yearly averaging periods.  Additionally, a very tentative



effort to construct an annual average nitrate transfer matrix



has been attempted.



     The purpose of this Phase II report is to provide as



complete a response as is currently as possible to all the



scientific and technical areas identified in the Terms of



Reference and as specified in the approved work plan of Work



Group 2.

-------
                             1-3'


1.2.1  Modeling Subgroup

     During Phase II/ the Modeling Subgroup of Work Group 2

has devoted its efforts to:

(1)  a synthesis of the Phase I transfer matrices into a
     "best estimate" for preliminary assessment iterations

(2)  evaluation of selected sulfur regional transport models
     against monitoring data for ambient concentrations and
     deposition rates;

(3)  intercomparison of model results in order to analyze their
     performance characteristics and further improve their
     reliability;

(4)  production of the Phase II transfer matrices using
     improved Phase II input data bases, and identification
     and analysis of the variations and uncertainties in the
     individual matrix elements;

(5)  production of preliminary transfer matrices for nitrogen
     oxides using several simplified NOX-N03 chemistries in
     selected models;

(6)  preparation for Phase III activities, including review
     of the regionalization for source areas and of emissions
     data bases.

A comprehensive report on the efforts of this Subgroup and

separate reports on its several activities have been prepared

(see Appendix 5).

   1.2.2  Atmosheric Sciences Review and Monitoring Data
          Analysis Subgroups	

     Many advances in understanding the regional and long-

term transport of air pollutants have been gained in recent

years, in large part due to an expansion of research efforts

on the underlying physical phenonmena involved in transport,

transformation and deposition.  These efforts have driven

the development and use of large mathematical models to

-------
                             1-4.





integrate the scientific information as it becomes available.



Even so, it is not possible to describe fully in a practical



model all aspects of air pollution transport on a regional



or continental scale.  Consequently, many simplifications



have been made in the modeling analyses presented and dis-



cussed in this report.  The role of the Atmospheric Sciences



Review effort is to provide scientific analyses in selected



areas of basic atmospheric processes, so as to specify more-



precisely the validity and range of uncertainty that character-



ize the modeling methodologies utilized and presented in



this and subsequent reports.  A separate report on the Phase II



efforts of the Atmospheric Science Review Subgroup has been



prepared (see Appendix 5).



     Similarly, a Monitoring Data Analysis Subgroup has been



constituted to analyze and interpret available monitoring data



in order to gain further insights into transboundary air



pollution phenomena.  The initial work effort of the Monitoring



Data Analysis Subgroup is presented in Chapter 3 of this



report.  The Monitoring Data Analysis Subgroup (to be called



the Monitoring Interpretation Subgroup in Phase III) will



produce a separate report in Phase III.



1.3  Phase II Report





     This report is structured to follow closely the terms of



reference and work plan for the Group.  The following chapter

-------
                             1-5.

contains summaries of the four atmospheric science issues reviewed
during Phase II.  Chapter 3 contains an analysis of  monitoring
results and their interpretation.  The next two chapters
describe the role of models in the particular application at
hand, and the status of those models which have been selected
for use in Canada and the United States.  The Phase II improved
inventories for current S02 emissions in eastern Canada
and the U.S. at the state/province level and at the Acid
Rain Mitigation Study (ARMS) area level and for the top 50
S02 emitters in North America are presented in Chapter 6.
presented.  The seventh chapter presents the Phase II source-
receptor transfer matrices for sulfur oxides and compares them
to those for Phase I.  Chapter 8 contains a preliminary analyses
of the variations and uncertainties in the transfer matrices for
sulfur oxides.

     Chapter 9 and 10 contain preliminary transfer matrices
for nitrogen oxides and primary sulfates, respectively.
Chapter 11, "Conclusions, Recommendations, and Phase III
Work Plan", charts the future course of action for the Work
Group.  The final chapter contains preliminary proposals that
should be considered for inclusion within the Research, Modeling,
and Monitoring element of a transboundary agreement.

-------
                             1-6.


1.4  Phase III Activities

During Phase III, the Work Group will give additional

emphasis to:

(1)  expanding in scope and depth the Atmospheric Science
     Review and Monitoring Data Analysis efforts;

(2)  exercising the selected models using the Phase II improved
     SC>2 emission inventory on a state/province basis and
     attempting to unify the transfer matrices;

(3)  examining and quantifying the "background contributions"
     to concentrations and depositions in the targeted sensitive
     areas and the ventillation out of the region, particularly
     to north and east;

(4)  preparing the 1979 data bases for the second and third
     rounds of model evaluation;

(5)  exercising the selected models on additional periods of
     meteorological data to assess the seasonal and annual
     variabilities in source receptor-relationships;

(6)  preparing and implementing a detailed work plan for modeling
     additional transboundary air pollution issues; and

(7)  addressing the issues raised by the peer and other external
     reviews.

-------
                          Chapter 2


          ATMOSPHERIC SCIENCE ISSUES IN LRT MODELING

2.1  Introduction




     When nearing the completion of Phase I activities and
                                                        v
preparing the Phase I report, Work Group 2 held a workshop in

Washington, D.C. on December 16, 1980.  A wide ranging

discussion occurred regarding the most important areas in the

atmospheric sciences which were closely related to the use of

long-range transport models.  From these dicussions four

topics were chosen as requiring highest priority for immediate

review in the Phase II activities.  Designated members of the

Work Group were assigned the task of preparing review papers

with the help of outside experts.  These review papers were

to highlight the state of knowledge in the particular topic

area and to indicate how that knowledge is reflected in the

various models being used by the Work Group.

     The four review papers are presented in full in a

background document (Report No. 2-14) by the Atmospheric

Science Subgroup of Work Group 2.  A brief summary of the

most important and relevant (to modeling) findings of these

reviews follows.

2.2  Sulfur and Nitrogen Chemistry in LRT Models

     Present understanding of the homogeneous gas phase

reactions of SC>2 indicates that the rate of S02 oxidation in

the atmosphere is dominated by free radical reaction processes.

-------
                             2-2.




The free radical species identified as important contributors


to the SC>2 oxidation process are hydroxyl (HO), methylperoxyl


(CH302) and other organic peroxyl species (RC>2/ R'C>2 etc.).


The concentration of these radicals in the atmosphere are


dependent on many factors/ the more important of which are


the concentration of volatile organic compounds and nitrogen


oxides (NO and NO2) in the atmosphere, temperature and solar


intensity.  Theoretical estimates have shown that maximum S02


oxidation rates of 4.0% h~l are possible in polluted atmospheres.


However, recent experimental rate constant determinations for


the H02 and CH302 reactions with S02 indicate that these


processes may not be as important as previously thought and


that the maximum possible homogeneous S02 oxidation rate


under optimum atmospheric conditions may only be of the order


of 1.5% h"1.  This rate is the result of the reaction of S02


with the hydroxyl radical only.


     Present knowledge of heterogeneous pathways to S02


oxidation in the atmosphere indicates that the liquid phase

                                i 2
catalyzed oxidation of S02 by Mn   ion and carbon are


potentially important processes, as is oxidation by hydrogen


peroxide. Theoretical estimates of the maximum rate of


atmospheric S02 oxidation via these processes are of the


order of 10% h~l.  Unfortunately,  a great deal of uncertainty


surrounds the actual availability of these catalyzing substances


in ambient fine particulate matter.  The quantitative determination

-------
                             2-3
of rates of SC>2 oxidation via these processes has never been
demonstrated under actual atmospheric conditions.
     Organic and nitrate particulate matter forming processes
are presently thought to be dominated by homogeneous gas
phase reactions.  In the case of atmospheric nitrates, a
particularly significant production pathway is through a reaction
between the hydroxyl free radical and nitrogen dioxide resulting
in nitric acid  (HON02) formation.  The fate of nitric acid in
the atmosphere  is not well understood, though a portion of
gaseous nitric acid is known to enter into an equilibrium
with ammonia (NH3) to form particulate ammonium nitrate
(NH4N03).  Present knowledge provides little support for
liquid phase oxidation as an important pathway to NOX
transformation.
2.3  Trends in Precipitation Composition and Deposition
     Establishing trends in the chemical concentrations in
precipitation or the deposition of these materials in eastern
North America over the past 20 to 30 years is difficult
because the appropriate continuous data sets do not.exist.
Various stations or networks have been established by different
agencies for different purposes at different times during
this period.  However, no one network has been operated for
more than a few years, and any trend must be established by
interpolation and extrapolation of the existing data sets.

-------
                             2-4






All the factors that can contribute to the difficultly in doing



this are pointed out in Chapter 2 of Report No. 2-14.  Notwith-



standing these difficulties, the data do suggest an expansion



of the region covered by acidic rainfall, especially into



the southern and western U.S.  There are differences of opinion



on whether the acidic deposition has actually increased in



amount, although on balance the data support the continuation



of elevated acidity levels at many locations in the northeastern



U.S. and eastern Canada.



     Its becomes even more difficult to relate the suggested



trends to changes in emissions because the emissions were



not nearly so well documented in the 1950s and 1960s before



the statutory requirements of the clean air legislation



in the U.S.  and Canada.  The best estimates indicate a



moderate increase (approximately 20-40%) for SC>2, and a



dramatic increase (approximately 300%) for NOX emissions



between 1940 and the mid-1970s in the U.S.  However, other



characteristics of the emissions have also changed; such as



the removal of a greater portion of the particulate loading



since 1971 and the steadily increasing emission heights in



recent years.



     The Modeling Subgroup has focused its modeling efforts



on the present day situation using the best current estimates



of emissions and evaluation against the current multi-network



deposition measurements.  These models are not being applied



against past data because of the uncertainties in past measure-

-------
                             2-5


merits and input data.  Thus, trends are a non-issue as far as

the Work Group 2 modeling is concerned.  However, it is

important to better establish the existence (or non-existence)

of past trends and to understand their significance and,

therefore, further work needs to be carried out.  As for the

future, a strong commitment by the appropriate agencies in

Canada and the U.S. to continuation of monitoring networks

is required to ensure that we will not be faced with a lack

of appropriate data 10 and 20 years from now.

2.4  Seasonal Dependence of Atmospheric Deposition and Chemical
     Transformation Rates for Sulfur and Nitrogen Compounds

     A literature survey has been carried out into the seasonal

variations of the wet and dry deposition rates, as well as

the chemical transformation rates, of sulfur and nitrogen

oxides, with particular reference to deposition and transformation

parameters of relevance to long-range transport models.  Both

relevant theoretical and experimental results have been

considered although a critical evaluation of the literature

has not been attempted.

     From a theoretical view point, the deposition and trans-

formation rates of sulfur and nitrogen compounds could

potentially have a substantial seasonal variation.  However,

it is difficult to draw conclusions about the magnitude of

this variation with any degree of confidence from the current

theories, with the possible exception of the wet and dry

-------
                             2^6

deposition of sulfur dioxide and the photochemical component
of its chemical transformation rate.  Therefore/ the available
field data were also considered, although these were often
too scanty to be of much assistance.
     An attempt was made to summarize the available information
on the seasonal variation of transformation/deposition rates
for the sulfur compounds.  It was not intended to recommend
these values for use by the long-range transport modeler -
much more experimental and theoretical work is needed before
this will be possible - but rather, to indicate whether
seasonal changes in the parameter of interest are expected to
be greater or less than an order of magnitude.  At present,
little more than this can be done.  The conclusions are as follows;
     1.  The scanty available data suggest that the washout
         rates of sulfates (and probably nitrates) should be
         comparable in summer and winter.  The rainout rates
         could be strongly dependent on storm type, and hence
         the time of year, because of the different mechanisms
         whereby particles can be incorporated into precipita-
         tion (  - some data suggest variations of an order-
         of-magnitude or more).
     2.  Experimental results and theoretical considerations
         suggest a seasonal variation of the wet scavenging
         coefficient for sulfur dioxide which can be up to
         several orders of magnitude, depending on the latitude.

-------
                        2-7

    This variation is most pronounced in the northern
    parts of America which receive appreciable amounts
    of snow in the winter and convective storms in the
    summer.  Probably the same conclusions also apply
    to nitrogen dioxide.  Nitric acid vapor, being
    highly reactive with all kinds of surfaces/ is
    expected to show a smaller seasonal dependence of
    the scavenging coefficient.
3.  The situation is too confusing at present to draw
    any conclusions about the seasonal dependence of
    the dry deposition rate for sulfates (or nitrates).
    In the winter/ deposition velocities would seem to
    be 0.2 cms~l or less, but values reported for summertime
    conditions range over an order of magnitude, including
    negative numbers.
4.  The dry deposition velocity of sulfur dioxide is
    expected, from available experimental and theoretical
    results, to show only a modest seasonal variation -
    generally, less than a factor-of-two or so in any
    given area.  The same is probably true of nitrogen
    dioxide and nitric acid vapor.
5.  The gas-phase homogeneous component of sulfuric and
    nitric acid formation rates is relatively well
    understood, and has a strong seasonal variability,
    especially at the northern latitudes.  However, our
    knowledge of the heterogeneous component, including

-------
                        2-8





    in-cloud processes/ is too poor at present to allow



    any conclusions regarding the seasonal dependence



    of the overall chemical transformation rate of



    sulfur and nitrogen oxides.



6.   For many of the parameters under consideration,



    during the winter months/ rates are strongly depen-



    dent on latitude - e.g., photochemical conversion



    rates of sulfur and nitrogen oxides above 45°N become



    negligible, as do also wet deposition rates of gases



    such as sulfur dioxide (because precipitation is largely



    in the form of dry snow).  This indicates that not



    only the seasonal, but also the spatial variability



    of deposition and transformation rates should be taken



    into account in long-range transport models.  Although



    it may be too early to speculate, the following



    approach does not seem unreasonable:  during the summer



    months, one might assume, as a first approximation,



    the same values for deposition/transformation parameters



    regardless of location, for each species of interest.



    During the winter months, while rates at the southerly



    latitudes might stay roughly the same as those in the



    summer/ the models would include a dependence of



    deposition/transformation on latitude/ which could



    be quite pronounced for some of the parameters (such



    as wet deposition of sulfur dioxide).

-------
                             2-9-


     7.  For the sulfur compounds, more experimental data

         are badly needed, both under summer and winter-

         time conditions, particularly on wet and dry

         deposition rates of particulates and chemical trans-

         formation rates in regional scale air masses (as

         opposed to chimney plumes).  Very little is also

         known about in-cloud transformation and deposition

         processes.  For the nitrogen compounds, data are

         required in almost every area of interest, and

         immediate support for laboratory and field investi-

         gations into deposition and transformation rates

         of the major species (NO, N02/ HNC>3, nitrates and

         PAN) is strongly recommended.

2.5  The Global Distribution of Acidic Precipitation and Its
     Implications for Eastern North America	

     A review of the world-wide data on precipitation pH in

remote and exposed mid-latitude west-coast areas indicates

that all precipitation contains acidic materials in small

quantity.  In the absence of any neutralizing alkaline

components, the acidic materials are sufficient to reduce the

pH to a value of about 5.0 and in some cases less.  Nowhere,

though, are pH values in remote areas as low as those found in

the most acidic precipitation areas of northeastern U.S. and

western Europe.  Several authors are now suggesting that the

reference level of 5.6 (the pH of rainwater in equilibrium

with atmospheric €62) is not appropriate and that departures

-------
                             2-1-0

from a value of near 5.0 would indicate the regional and local
modulations of the influence of "global background".
     While pH is a useful single number that characterizes
the precipitation, it is the total deposition of acidity (H+
ions) and sulfate that is important in assessing the effects on
sensitive ecosystems.  The deposition is the product of the
concentration and the rainfall amount.  Thus in considering
the relevance of the low pH values in remote areas, this
must be considered.  Sensitivity, in the form of the buffering
capacity of the receptor surfaces is also important in defining
the seriousness of impact.  Most remote areas, especially
arid regions, are well buffered and so the impact of any
acidic deposition is minimized.  In contrast, the regions
with lowest pH and highest depositions of H+ ion in the
northeastern U.S., eastern Canada and southern Scandinavia
cover large areas of poorly buffered lakes and soils and
thus have a major impact on the receptors there.
     While there is considerable variability in the background
pH values, they are, in general, consistent with the concepts
proposed.  The limited vertical profiles available also are
supportive of the hypothesis that most precipitation starts
off as acidic cloud droplets.  It must also be pointed out
that some of the observations cannot be readily explained and
clearly, more analysis of existing data bases are required.
For example, trajectory analyses could be used to identify

-------
                             2-1-1

whether observed background levels are due to natural sources
or residuals from man-made sources far upwind.
     The number of remote stations that have been established
                                             •9
in recent years is now beginning to generate substantial data
relevant to the issue of establishing and understanding the
background levels of air and precipitation chemistry.  Rather,
than establishing many more such stations, the priority should
be to analyze and interpret the existing data base.
     Some specific recommendations are as follows:
          wherever possible at precipitation chemistry
          stations, sampling should be done on an event
          or at least on a weekly basis.
          the precipitation chemistry data is much more
          valuable and can be interpreted more readily
          if concurrent basic air chemistry measurements
          are made by filter-pack sampling.
          more observations of the vertical distribution
          of precipitation chemistry (and where possible
          air chemistry) are needed.  This can be done in
          two ways
          a)  at mountain sites
          b)  with instrumented aircraft

-------
                   2-12


continuing efforts are required to refine the

estimates of natural emissions of acid compo-

nents into the atmosphere (they are presently
                  •»
less accurate than estimates of man-made

emissions/ yet are equally important on the

global scale).

estimates (however approximate) are required

for emissions of the most important alkaline

materials into the atmosphere; at present

none exist.

-------
                            Chapter 3

              MONITORING RESULTS AND INTERPRETATION
3*1  North American Precipitation Acidity
     Precipitation chemistry was monitored at approximately ninety
locations in North America by the governments of the United States
and Canada in 1979.  The observed annual mean precipitation-weighted
pH distribution is shown in Figure 3-1.  The annual-mean pH of
precipitation was lowest in the eastern half of the continent.
The lowest pH (highest acidity) of about 4.2 occurred in a corridor
stretching through Ohio and Pennsylvania into southern Ontario.
The zone of maximum acidity is immediately downwind of large
sources in the upper Ohio River Valley.  During periods of preci-
pitation, the prevailing flow is generally from the southwest in the
eastern U.S. and Canada.  This situation facilitates the transport
of pollutants from major sources to the northeastern U.S. and
southeastern Canada.
     A zone of high pH at mid-continent to the east of the Rocky
Mountains indicates the simultaneous absence of large acidic-
pollution sources in the region and the presence of alkaline
soil.  Therefore wind-blown dust acts to cause more basic rain.
This situation is only typical of dry mid-continental areas where
abundant sources of alkaline materials are available.
     On the west coast of the continent, precipitation is slightly
acidic (annual - value of about 5.6).  The cause of that acidity
is not well understood.  Possible explanations include:  anthro-
pogenic sources which do exist, the release of biologically-

-------
                                                                         .30"
^Figure  3-1.
Annual Average pH of Precipitation in North America
During 1979 Based On Observations By Canadian  APN
and CANSAP Networks and American MAP3S and NADP  Net-
works. (Note: An isoline is dashed where  uncertainties
in its position are great due to lack of  data.)
Issued by the Atmospheric Environment Service  May  1981.
            110
             100'
90"
j
                                                           HO*

-------
                               3-3

produced organic sulfur compounds from the Pacific Ocean surface
that ultimately oxidize to sulfur dioxide and sulfuric acid in
the absence of neutralizing materials.
     Further east/ the west coast mountains in North America
influence the continental pH distribution by acting as a cleansing
barrier.  Strong precipitation scavenging induced by orographic
lift causes the air mass impinging from the west to produce
precipitation with a pH close to 5.6 in the absence of alkaline
wind-blown dust.  This is borne out by observed pH in the vicinity
of the mountains of about 5.6.
     To give an overview of the hemispheric acidity pattern, an
estimated pH distribution is shown in Figure 3-2 (Gravenhorst,
et al. 1980).  Of note are the high precipitation acidity in
Eastern North America, Western Europe and Japan, and the more
alkaline precipitation in the large continental areas.
3.2  Temporal Variations
     Air
     In 1979, routine monitoring of the daily average concentration
of sulfur dioxide, sulfates and precipitation chemistry was
carried out at rural, regionally representaive sites (Figure
3-3) in eastern Canada (Barrie et al., 1980).  The network is
the Canadian Air and Precipitation Monitoring Network (APN).
     Monthly mean concentrations of sulfur dioxide and particulate
sulfate are given in Figure 3-4.  On an annual basis, the lowest

-------
                              3-4
Figure 3-2.    Estimated Distribution of the pH For Rain Water
               in the Northern Hemisphere. Source: Gravenhorst,
               et al.(1980).

-------
                                3-5
                                              APN
                                                        KEJIMKUJIK
     txPERIMENTA
     LAKFS AREA
                                           • PRESENTLY OPERATING

                                           O PLANNED
Figure  3-3.
Sites in  the  Canadian Air  and Precipitation
Monitoring  Network (APN) At  Which Daily  Air
and Precipitation Samples  Are Collected.

-------
                                  3-6
op


 I
O
-

I
O
    15 -
    10 -
                                                        LONG POINT
                                                        CHALK RIVER


                                                	KEJIMKUJIK


                                                	ELA-KENORA
SULPHUR DIOXIDE
                                                            N
   Figure  3-4,
         Temporal Variations of the Monthly Average Con-

         centration  of Atmospheric Sulfate and Sulfur  Di-

         oxide At APN  Sites During 1979.

-------
                               3.-7



concentrations occurred at Experimental Lakes Area (ELA)-Kenora

                                       -J                 -3
in northwestern Ontario (S02 = 1.5 jug/m ; S04 = 1.6 jjg m~°).


In the mid-latitudinal westerlies, ELA is upwind of major North


American sulphur sources.   Long Point, on the other hand is


located in an industrialized section of Ontario and is northeast


of major emissions in Ohio and Pennsylvania.  At all  sites in


eastern Canada, there are seasonal variations of atmospheric


sulfur dioxide levels.  The levels are highest in the winter and


lowest in the summer.  On a percentage basis the amount of variation


about the annual mean concentration depends on the location.  It


is lowest in source regions and highest at samplers which are remote


from source regions.  The percent standard deviation  about the


annual average concentration is 43 af Long Point on Lake Erie,


63 at Chalk River (500 km away), 75 at Kejimkujk and  108 at ELA,


the furthest location from the major source area of emission.


     Particulate sulfate concentrations are highest in summer at


all APN sites except ELA-Kenora.  A winter maximum in both sulfur


species at Kenora owes its existence largely to meteorological


factors.  Maximum transport westward from eastern North America


in winter coincides with a large winter peak of background sulfur


concentrations in artic air masses (Barrie et al., 1981) that


prevail at this mid-continental location during winter.


     At APN sites located in the continental pollution plume, the


sulfate seasonal cycle is one hundred eighty degrees  out-of-


phase with the sulfur dioxide cycle.  Indications are that the

-------
                               .3-8


summer sulfate maximum owes it existence to a summer maximum in

the conversion rate of sulfur dioxide to sulfate.  One

manifestation of higher S02 to sulfate conversion in summer is

that, at all stations/ the ratio of S02 to total airborne sulfur

(S02 + 304) tends to be lowest in summer (e.g. see Figure

3-5a for Long Point).

     On a daily basis, the concentrations of sulfur oxides are

highly episodic regardless of location in eastern Canada.  Polluted

and non-polluted periods of 3 - 6 day duration alternate regularly.

The dry deposition rates of these acidic substances, which is

thought to be roughly proportional to their atmospheric concentra-

tions, is equally episodic.  A comparison of monthly wet and dry

deposition rates of sulfur compounds is made later in this chapter.

     Recent work in the U.S. concerning transport of particulate

sulfate has been reported by Paresh and Husain (1981).  In this

study, particulate sulfate in air was monitored continuously from

June 1978 through December 1979 at Whiteface Mountain, New York.
                            =
The influx of transported 804 was evaluated by sectors to assess

the relative contribution from the U.S. and from Canada.  The

daily 304 concentrations were related to surface-air trajectory

ensembles.  During the study period the site was influenced

approximately equally by the Canadian continental polar winds and

the U.S. maritime tropical winds.  However, the maritime air masses

from the U.S. were the principal conveyors of very high urban-

like 304 concentrations at this site and transported about 4 to 5

-------
                                 3-9
       100 -
  Psl
 a
 CE
 cr
 Lu

 Ld
 _J
 a
 UJ
 _j
 a
!! IT
 cn
 LTl
 4-
  PJ
 a
 t_n
          0
     2.00 -
       .00
     0.00
             N   D

        Figure 3-5.
 j'r'H'n'M'J1 .j'n'51  D'N'D

Top: The Fraction of Total Airborne Sulfur Exist-
ing As Sulfur Dioxide At Long Point On the North
Shore of Lake Erie on a Daily Basis (1978-1979).
Bottom: The Concentration of Total Airborne Sulfur
At Long Point On a Daily Basis. (Note the  preval-
ence of episodes of  elevated sulfur levels of 3-6
davs duration.)

-------
                               3-10
times more 804 than did the continental polar air masses from
Canada.
     Precipitation
     The temporal variations of monthly precipitation-weighted
mean concentrations of the four major ions (H+, NH4+, S04 and N03~)
at APN sites during 1979 are shown in Figure 3-6.
     At sites nearest to the major pollutant source region (i.e.,
Long Point and Chalk River), sulfate concentrations were highest in
the summer half of the year and lowest in the winter half, while
hydrogen and nitrate ion concentrations showed no significant
seasonal variations.  A similar seasonal variation of sulfate
and nitrate in precipitation has been observed south of the border
in the American northeast in the MAP3S network(MAP3S/RAINE,1981)
(see Figure 3-7).  However, the seasonal variation of hydrogen ion
follows the sulfate in the MAP3S data.  At ELA-Kenora, a site more
remote from the influence of continental sources, sulfate did not
peak in any particular season, nitrate concentrations varied little,
but precipitation acidity was highest in winter.  The latter was
accompanied by a minimum in ammonium ion concentration and also
coincides with the time of year when wind-blown dust would be a
minimum.  On an equivalent basis, sulfate was more abundant than
nitrate throughout the year at all stations.   An exception was
at Chalk River when they were equal in concentration during the
winter.

-------
    400
    200
           LONG POINT
                              3-11
                                                      H+
                                                ------ NO;
                                                          T	1
    200
 ^
 cr

 3  100
 CO
m   0
Z
o

S  BO
z
o
          CHALK RIVER
                                                           i	I
LU

>

X

Z
O
     40
      0
           KEJIMKUJIK
Figure 3-6,
                                                          N
              The Temporal  Variation of Monthly Precipitation-
              Amount Weighted Mean Concentrations of Four Major
              Ions At  APN Sites During 1979  (see Fig.  3-3).

-------
                                   3-12.

              /u+\
   8-1
 O
 o
 c
 o
 O
      X
      JASON DUFMAMJJASON DU F  M A M  J
          1977                1978                 1979
   s-
              (S04=)
                                                                LEGEND
                                                              = Wh!t«face
                                                                Ithaca
                                                           .  + = P«nn Stat«
                                                             x = Vlrglnlo
                               A.                             Os III! no I s
                              .'+;.   .O _                    *  V = 8rookhav«n
         S. .x               «  .'.'«'••*.•  ••                   •'   B = D«lawar«
                          ______          __x = 0hlo
 O
 O

   o
      JASONDUFMAMJJASONDIJFMAMJ
          1977                1978                 1979
   O-,
   oo
o
C o
                (N03  )
      v                       A"  •' '•      7  ^  '-
      X.  -f X                  .»-._• * •    n • '  *  A-
O
 ••«-*•
O
O
                        "0
                                     1   I  r  T   i  T  I   r
      JASONDIJFMAM  J  JASONDlJFMAM J
           1977                   1978                   1979


 FIGURE 3-7  Monthly Mean Deposition-Weighted  Concentrations for
             H+,  S04 and    ~

-------
                               3-13





     Wet Versus Dry Deposition



     Monthly wet and dry deposition of sulfur oxides were



derived from APN air and precipitation data (see Figure 3-8).



Dry depositions was calculated from air concentrations measured



by the APN network (Figure 3-2) and from the deposition velocities



listed in Tables 3-1 and 3-2.  Deposition velocities were estimated



using monthly frequency distributions of Pasquill-Gifford stability



classes and the techniques of Sheih et al.  (1979), modified to



include higher particulate sulfate surface-resistances.  These



values can be considered only rough estimates of the dry deposition,



Wet deposition was calculated from the product of precipitation



amount and ionic concentrations (Figure 3-6).



     In general, dry deposition of sulfur is not negligible



compared to wet.  The ratio of dry to wet deposition is highest



for locations that are close to source regions.  It is also



higher in winter than in summer since in winter, SC>2 concentrations



near the ground are highest and precipitation amount is usually



lowest.



3.3  Classification According to Air Parcel Origin



     Measurements of daily mean concentrations of particulate



major ions and sulfur dioxide made at APN sites (Figure 3-2) in



1979 were classified according to air parcel origin.



     Back-trajectories at a level of 925 mb were calculated for



every site and for four times (0, 6, 12, 18 'GMT) each day for a



period of five days based on six hour time steps within the

-------
          LONG POINT SO,
                    3-14
                                \
         j	  i
                            I	  J	I	I	I	I
 C\J

  E
  o
 O
 O
 Q.
 LU

 Q
 cc
 D

 Q.


 CO
  z
  o
     4r-
          CHALK RIVER
         X
             X
           x
                             J	I
                                      I	I
KEJIMKUJIK
       r
      1 -
Figure 3-8,
                                                          N
                                                     D
     The  Temporal Variation of Monthy Wet and  Dry  Dep-

     osition  of  Oxides of Sulfur At APN Sites  (see Fig.
     3-3).  Dry Deposition is Calculated From Air Concen-
     trations (Fig.  3-4)  and Deposition Velocities  (Tables
     3-1  and  3-2). Wet Deposition is Measured.

-------
                                         3-15
                  Table 3-1.   Estimated Monthly Average Dry Deposition
                               Velocities (cm s~l) of Sulfur Dioxide
                      	at AEN Sites
 SITE         JFMAMJJASOND
 ELA          0.15  0.15  0.45  0.43  0.43  0.43  0.42  0.42  0.41  0.43  0.43  0.15

 Long Point   0.15  0.15  0.71  0.68  0.64  0.41  0.37  0.37  0.55  0.58  0.65  0.15

 Chalk River  0.18  0.18  0.29  0.27  0.22  0.22  0.18  0.27  0.22  0.22  0.23  0.18

 Kejimkujik   0.18  0.18  0.44  0.37  0.37  0.36  0.30  0.30  0.25  0.28  0.29  0.17
                  Table 3-2.   Estimated Monthly Average Dry Deposition
                               Velocities (cm s"1) of Particulate Sulfate
                               at APN sites.
                                        MCNTH

SITE         JFMAMJJASOND


ELA          0.08  0.08  0.19  0.19  0.20  0.15  0.15  0.15  0.14  0.15  0.14  0.08

Long Point   0.08  0.08  0.10  0.10  0.09  0.08  0.08  0.08  0.10  0.10  0.11  0.08

Chalk River  0.10  0.10  0.32  0.29  0.27  0.30  0.30  0.29  0.18  0.18  0.19  0.10

Kejimkujik   0.09  0.09  0.33  0.33  0.34  0.24  0.25  0.23  0.17  0.17  0.17  0.09

-------
                               3-16





5-day period.  However, only the last two days of the trajectories



were used in the classifications since most major sources are



within two days travel time of the locations under consideration



and the accuracy of trajectories decreases after two days.



     For this preliminary sector analysis, five concentration



ranges were established for each chemical species in air and



precipitation.  The compass was divided into eight equal sectors.



Four 2-day back-trajectories were used to determine the sector froi



which most of the air originated in a 24-hour period.  Each



observed daily concentration was classified into a particular



concentration range (Table 3-3) and compass sector.



     The trajectory-wind-roses at APN sites were plotted on



a map of North America (Figures 3-9 to 3-13).  There is a differenl



map for each chemical species.  In addition to the frequency



of occurrence of air from any sector being represented by the



length of the radial projection for each width, the wider the



projection the higher the concentration of each chemical species.



Concentration ranges are listed in Table 3-3  for particulate



sulfate, sulfur dioxide, pH, precipitation sulfate and precipita-



tion nitrate.



     The reader is cautioned that the reported trajectory-wind-



roses for precipitation days (Figures 3-11 to 3-13) may contain



uncertainties.  Precipitation events are often accompanied by



frontal passages, the presence of which can reduce the accuracy



of an air parcel trajectory analysis.  Nevertheless, it is felt

-------
                   3-17
Table 3-3   Definition of Concentration and pH Ranges
  '    	for the Trajectory Cases	

Particulate Sulfate : bdl* -
2 -
4 -
10 -
>
Gaseous Sulfur Dioxide : bdl
3 -
6 -
15 -
>
Rain pH : 3.80
(Note that in Figure 3-11 the 3.80 -
lower pH ranges correspond 4.40 -
to wider radial projections) 5.00 -
Rain Sulfate : bdl -
2 -
4 -
8 -
>
Rain Nitrate : bdl
1.29 -
2.58 -
5.17 -
*below detection limit >

2 jug/m3
4
10
20
20
3 /ag/m3
6
15
30
30
4.40
5.00
5.60
5.60
2 mg/1
4
8
16
16
1.29 mg/1
2.58
5.17
10.33
10.33

-------
Figure 3-9,
Trajectory Wind Roses At the APN Sites  For  Part-
iculate Sulfate Based On Two Day Back Trajectories
At 925 mb(about 800 meters) During January  -  Dec-
ember 1979. (Scales: 9mm = 10% and see  Table  3-3)
                                                                                         oo
                                  M>   o „•$;,•      o^u  \

-------
Figure 3-10.    Trajectory Wind Roses AtlH^N Sites For Sulfur Dioxide
               Based On Two Day Back Trajectories At 925 mb(about
               800 meters) During January - December 1979. (Scales:
               9mm = 10% and see Table 3-3)
                                                                                         Ul
                                                                                         I

-------
Figure 3-11.
Trajectory Wind Roses  At^^N Sites For Rain pH Based
On Two Day Back Trajectories At  925 mb(about 800
meters) During January - December 1979.  (Scales: 9mm
= 10% and see Table 3-3)

•"       '    :tr^

-------
Figure 3-12.    Trajectory Wind Roses At^B>N Sites For Rain Sulfate
               Based On Two Day Back Trajectories At 925 mb(about
               800 meters) During January - December 1979. (Scales:
               9mm = 10% and see Table 3-3)
             *-	(	:»"3

-------
Figure 3-13.
Trajectory Wind Roses  At^PPN  Sites  For Rain Nitrate

Based On Two Day Back  Trajectories  At 925mb(about
800 meters) During  January -  December 1979. (Scales:
9mm = 10% and see Table  3-3)
                                                                                       U)
                                                                                       I
                                                                                       ro
                                                                                       NJ
                          r^>^>r//'o r(* •«
                          \'~t• JL'H/-  -I'-o'


-------
                               3-23


that the graphs in Figures 3-11 to 3-13 are accurate enough to

allow one to draw some valuable conclusions.

     The trajectory-wind-rose data contain a great deal of

information pertinent to the long-range transport of pollutants.

Some of the most important results are summarized as follows:

(1)  in general/ the highest concentrations of acid-related
     substances in air or precipitation occur when air originates
     from a sector between southeast and northwest; the lowest for
     air from a sector between north and southeast;

(2)  southerly and southwesterly trajectories are more prevalent
     during precipitation periods than during the whole year
     (compare the envelopes of trajectories in Figures 3-7 and 3-9);

(3)  the phenomenon of long-range transport is clearly demonstrated
     by results at Kejimkujik National Park in Nova Scotia which
     is downind of and remote from continental sources.  The highest
     concentrations in air and precipitation occur mostly when air
     originated from the southwest to west sector.

     In the U.S., the Air Resources Laboratories' Atmospheric

Transport and Dispersion Model (ARL-ATAD) has been used to develop

a trajectory climatology in conjunction with the precipitation

chemistry at MAP3S Whiteface Mountain and Illinois sites (Wilson,

et al., 1980).  Trajectories were calculated for all of 1978 and

1979 and weighted by precipitation amounts during every six-hour

period to determine a dominant direction or origin of air mass

for each precipitation event.  Each event was then classified in

one of twelve sectors covering 30° each.  Only the final two

days (48 hours) of approach trajectory end points were considered.

This allowed differentiation between so-called "Ohio Valley/Midwest"

and "Canadian/Great Lakes" air masses.  The so-classified wet

ion deposition as a function of trajectory sector are shown in

Figures 3-14 through 3-16.  The Ohio Valley/Midwest events were

-------
                                3-24
                                OHIO VALLEY/MIDWEST-

                                CANADIAN/GREAT LAKES-

                                TOTAL PRECIPITATION-
                                                   17.3  LITER

                                                    7.9  LITER

                                                   30.6  LITER
                                                     26X OF TOTAL.
           ee
 T
ee
                    1 2Q
                              i se  i eo  21 a 243  270
                                 TRAJECTORY SECTOR
                                                              see
b)
te -
                                OHIO  VALLEY/MIDWEST-

                                CANADIAN/GREAT LAKES-

                                TOTAL DEPOSITION-
                                                   23.3  MG/M2

                                                   11.7  MG/M2

                                                   37 . 4  MG/M2
                                                   3IX OP TOTAL.
                                      OHIO  VAL.L.EY/

                                        MXOWEST
                                                   CANADIAN/

                                                  GREAT  UAKCS
                              tee  i eo  2 IB 2-*e  270
                                 TRAOECTORY SECTOR
                                               see 338  see
Figure 3-14.
          Top:  Precipitation Volume, and Bottom: Hydrogen
          Ion  Total  Wet Deposition, At Whiteface Mountain,
          New  York,  Based On Two Day Back Trajectories In
          the  Boundary Layer Assigned To 30° Sectors.
          Source:  Wilson, et al, 1980.

-------
  eeei
  see-
A

N

£
\
O

£
Z see
o
H
h
H

8
0. aecr
UJ
O
o
w i ee-
u
                 3-25

               OHIO VALLEY/MIDWEST-


               CANADIAN/GREAT LAKES-


               TOTAL DEPOSITION-
                                    1087.9 MG/M2


                                     531.6 MG/M2


                                    1700.5 MG/M2
              64 X OF  TOTALv
                                    3IX OF TOTAL.
 I


60
 I


98
     I     1    T


120  150  180 210  248  278

       TRAJECTORY  SECTOR
                                                  308  338  360
                             OHIO  VALLEY/MIDUEST-


                             CANADIA'N/GREAT  LAKES-


                             TOTAL DEPOSITJON-






                                   66X OF TOTAL.
                                         646.7  MG/M2


                                         283.4  MG/M2


                                         994.6 MG/M2
                                                 28 X OF TOTAL.
•Figure 3-15.
                                   CANADIAN/

                                  GREAT LAKES"
                 .    x    x    x    y
              T——r

              \ 6O  ISO  210 2-40  270

                TRAJECTORY SECTOR
                                                  300 330  aea
  Top: Sulfate,  and Bottom: Nitrate Total Wet  Dep-

  osition,  At  Whiteface Mountain, New York, Based

  On  Two  Day Back Trajectories In the Boundary Layer

  Assigned  To  30° Sectors. Source: Wilson, et  al,  1980,

-------
      ee ~\
      •4B
X 30
O
I
s/


0
M
     v>
     O
     Q.
     Ul
     Q
     O
     u
                           3-26


                            OHIO VALLEY/MIDUEST-

                            CANADIAN/GREAT LAKES-

                            TOTAL DEPOSITION-
                         Oe.3  MO/M2


                         17.7  MOXM2


                         83.7  MO/M2
                                      72X OF TOTAL
                                                         OF TOTAL
                  ~f

                 60
                      122
tea  tea  210
  TRAJECTORY
 T	T

24Q  27Q
SECTOR
                                                 sea  330  see
                                 OHIO VALLEY/MIDWEST-


                                 CANADIAN/GREAT LAKES-


                                 TOTAL DEPOSITXON-
                                          66X OF TOTAL
                                                   1 ee . e

                                                    ee.3

                                                   1 8-4 . «*
                                                     CANADIAN/

                                                    GREAT LAKES
                                                          MG/M2
                                                     37* OF TOTAL
                               t se  i ee  21 o
                                  TRAJECTORY
                                                      330  360
Figure 3-16.
           Top:  Calcium,  and Bottom: Ammonium Total Wet Dep-
           osition,  At Whiteface Mountain, New York, Based On
           Two Day Back Trajectories In the Boundary Layer
           Assigned  To 30° Sectors. Source: Wilson, et al, 1980

-------
                               3-27
defined as those having their origin in the sector 165° through
285°, while the Canadian/Great Lakes events were designated by
the sector 285°-030°.  The wet deposition directional patterns
for all ions were found to closely follow the precipitation
variability in every case.  The wet deposition cumulative total
of the Ohio Valley/Midwest sector was found to account for approx-
imately 50-70% of the total wet deposition that could be classified
into the 48-hour approach trajectories.  Wet deposition from the
Canadian/Great Lakes sector resulted largely from winter coastal
storms in the form of snow or mixed precipitation.  Even though
ionic concentrations were generally lower for these events/ the
large values or precipitation sample volume resulted in almost
30% of the t'otal wet deposition at Whiteface Mountain.
     The same analysis procedure (ARL-ATAD model) was applied to
the MAP3S precipitation chemistry for the Illinois site during
1978.  Since large pollutant sources for S02, and NOX lie to
the east or southeast of the Illinois stations, one might expect
to see a significant difference in the deposition pattern pollution-
related ions observed at Whiteface Mountain during the same year.
The cumulative wet deposition total at Illinois for these events
is plotted in Figures 3-17 and 3-18 revealing a very interesting
picture.  Approximately 70% of the total precipitation for Illinois
was found to be associated with the southwest approach sector.
This is consistent with synoptic considerations whereby southwesterly

-------
  o -
     7  ~
    K
    LJ
     s  -
    z
    0
    H
    t-
    <
    t-
    M
    a.
    H
    0
    LI
    a
    a.
  s -
                               3-28


                               SOUTHWEST SECTOR-

                               TOTAL. PRECIPITATION-
                                                  IS. SI LITI-TW

                                                  25.31 LITFIR
                         1 20
                          1	T	1	r

                          150  163  210  24O  270
                            TRAJECTORY  SECTOR
                                                  300  33Q  3(
     •4B -1
     35
                               SOUTHWEST SECTOR-

                               TOTAL. DEPOSITION-
                                                  35 . I 3 MG/T-.2

                                                  44.85 MG/M2
                                    78X OF TOTAL
    O
    I
    0
    H
O
0.
u
0
n
I
     10
                    oa
                              GO  I «W 2IO

                               Tl'AOt-'CTOKY
                                                           3eo
                                             OR
Figure 3-17,
            Top: Precipitation  Volume,  and Bottom: Hydrogen
            Ion Total Wet Deposition,  At Whiteface Mountain,
            New York, Based On  Two Pay Back Trajectories  In
            the Boundary Layer  Assigned To 30° Sectors. Source
            Wilson, et al, 1980.

-------
                               3-29
        soon
        4OCT
      f\
      N
      I
      o seer
      I
      0
      H
      H
      o
      0.
      u
      0
      8 i**i
      2
                                 SOUTHUEST SECTOR-
                                 TOTAL DEPOSITION-
                  64X Of TOTAL
                                       574.22 MG/M2
                                       903.04 MG/M2
              30
                           I 2 Q
                               1 60  180 210  240  270
                                 TRAJECTORY SECTOR
                                                    30O 330  363
                                 SOUTHWEST SECTOR-
                                 TOTAL DEPOSITION-
                                      I 1 OA . 4 I  MGXM'O
                                          .11  MG/M2
                                              OF TOTAL
Figure 3-18
Top: Sulfate, and Bottom:  Nitrate  Total Wet Dep-
osition/ At Whiteface Mountain,  New  York,  Based
On Two Day Back Trajectories  In  the  Boundary Layer
Assigned To 30° Sectors. Source: Wilson,  et al, 1980,

-------
                               3-30


flow normally precedes an approaching cyclone system and subsequent

deposition events.  The reminder of the precipitation was found

to be essentially equally distributed over the various sectors

with the exception of the 0-30° segment.

     Air masses approaching the Illinois site from the southwest

were found to deliver about 70% of the total annual precipitation

resulting in the deposition of approximately 70% of the major

ions (Figures 3-17 and 3-18).  The analysis of precipitation

chemistry by air mass at both sites showed:


                                Whiteface Mountain (1978)

Midwest/Ohio Valley        56% of the annual precipitation delivering;
(160°-280° sector)

                           62% of the annual [H+] deposition

                           64% of the annual [864] deposition, and

                           65% of the annual [NC>3~] deposition

Canadian/Great Lakes       26% of the annual precipitation delivering:
(280°-030° sector)
                           31% of the annual [H"1"] deposition

                           31% of the annual [804] deposition, and

                           28% of the annual [N03~] deposition

                                        Illinois (1978)

(160°-280° sector)         71% of the annual precipitation delivering:

                           78% of the annual [H+] deposition

                           67% of the annual [804] deposition, and

                           64% of the annual [NO-j"] deposition

     Wilson, et al. (1980) concluded the following:

-------
                               3.-31





     (1) precipitation volume, more than any other single factor/



determines the amount of deposition for the three pollution-related



ions,  [H+] ,  [S04=] and [N03~]; (2) deposition of ions from



the Midwest/Ohio Valley air and Great Lakes/Canadian air does



not reflect the very significant differences in emissions that



are located in these two regions; (3) the Illinois results further



substantiate the above conclusion; (4) there seems to exist no



simple, straightforward relationship between emission source(s)



for acid precipitation precursor gas(es) and receptors of "acid



rain," i.e., [H+], [S04=] and  [N03~]  ions; (5) the chemical



transformation pathway(s) seem to be complex and insensitive to a



48-hour back trajectory analysis presented here and based upon



available meteorological and chemical information; and (6) the



concept of a "superbowl" (synoptic scale air shed) would explain



to some extent the uniformity in directional variability of



annual wet ion deposition in that the final products, i.e.,



acidic material is being rather evently deposited over a very



large region.



     These conclusions should be checked at Canadian and other



eastern U.S. monitoring sites.  A summary of the major source-



receptor analyses using trajectories or storm tracks is provided



in Table 3-4.

-------
                                       3-32
Table 3-4.  Major Source - Receptor Analyses Using Trajectories or Storm Tracks
PRINCIPAL AUTHOR
POLLUTANT
LOCATION
          CONCLUSION
SAMSON (1980)
S04
WHITEFACE    STRONGLY CORRELATED WITH SOUTH-
MOUNTAIN     WEST WINDS AND STAGNATION _> 36
             HOURS BEFORE UPWIND
WILSON, ET AL,.
(1980)
H+ AND
WET SO,
WHITEFACE    STRONGLY CONTROLLED BY PRECIPI-
MOUNTAIN     TATION AMOUNT AND CORRELATED
             WITH SOUTHWEST WINDS
PAREKH AND HUSAIN     804
(1981)
              WHITEFACE    MARITIME TROPICAL WINDS (FROM
              MOUNTAIN     U.S.) CONTRIBUTE 4-5 TIMES MORE
                           THAN CONTINENTAL POLAR WINDS
                           (FROM CANADA)
BARRIE, ET AL.,
(1981)
AND WET S04
AND WET N03
6 SITES      GENERALLY HIGHEST VALUES FOR
IN EASTERN   TRAJECTORIES FROM THE SOUTH-
CANADA       EAST TO NORTHWEST SECTOR
MERRITT  (1976)
40 TRACE
ELEMENTS
CHALK RIVER
 (CANADA)
GENERALLY HIGHEST CONCENTRA-
TION FOR STORMS TRACKS FROM
THE SOUTHWEST

-------
                               3-33



3.4  Episodes of Lorig-Rarige transport


     Episodes of elevated pollutant concentrations over a regional


area have been documented in several studies (c.f. Whelpdale, 1978;


Chung, 1977; Mueller et al.,*1979; and Lyons et al./ 1980).  One


of the most common meteorological situations responsible for


regional episodes is the large slow-moving high pressure system


or anticyclone. Over eastern North America, these occur in summer


or winter.  They usually contain extremely hot air in the summer


and very cold air in the winter.  Consequently, pollutant emissions


related to either air conditioning or space heating are high.


Typically, a low level inversion develops that traps pollutants


in the lowest 1-2 km of the atmosphere, thereby preventing their
                 •

dispersal.  In summer, the intensity of solar radiation is


sufficient to cause considerable photochemical activity in the


relatively cloud-free high pressure area and leads to elevated


ozone concentrations.  A particular episode is described in


detail below.


     In February 1979, a winter regional pollution episode


occurred in eastern North America in conjunction with a cold


high pressure system. It produced extremely high concentrations


of sulfur dioxide and particulate sulfates in eastern Canada.


Regional pollutant levels in the air mass entering southwestern


Ontario were so high that industries were requested by the Ontario


government to reduce emissions so as to not exacerbate the problem.


The episode was monitored in eastern Canada at rural sites in the

-------
                               3-34





APN Network (Figure 3-2) by Environment Canada and in southern



Ontario by the Ontario Ministry of Environment.  Reports of the



incident were published independently by Barrie et al. (1980) and



Shenfeld et: a 1. (1980).



     The cold air mass comprising the high pressure zone originated



in northwestern Canada.  The center of the high pressure zone



moved from mid-continent progressively eastward until it was in



the position shown in Figure 3-19 and 20 on February 19 and 20,



respectively.  The air mass was very stable with an intense



noctural ground-based inversion and overruning warmer air aloft.



     In eastern Canada elevated concentrations of oxides of



sulfur (Figure 3-21) occurred first at ELA-Kenora in northwestern



Ontario on February 17 as the centre of the high pressure zone



passed east of the site.  As the high pressure zone approached,



daily mean pollutant concentrations increased from relatively



low values (0.76 jug m   S04= and 1.1 jug m~3 S02) to maximum



values on February 18-19 (4-9 pg m~3 S04= and 18-29 /ag m~3



S02) as the southerly flow to the west of the center of the



high (see Figure 3-19) carried pollutants from the lower Great



Lakes region a distance of over 1000 km to the site.  Air parcel



back-trajectories are shown in Figure 3-22.



     In southern Ontario, the pollution episode peaked on February



20, one day later than in northwestern Ontario.  At this time,



the high pressure zone was positioned on the east coast (Figure



3-20) giving rise to a southerly flow over eastern North America.

-------
Figure 3-19.   Surface Weather Map  Sho^Wg Large High Pressure
               Area Centered Over the  Midwestern States On
               February 19, 1979.
1000  x IQI2 1016                        ,
                                                                                          u>
                                                                                           I
                                                                                          GO
                                                                                          Ul
                                                     io?:o
                                                                   GV.

-------
      Figure  3-20.
1012
1015
                 1020
Surface Weather Map  Showing the Movement of  the
Large High Pressure  Area From the Midwest  (see
Figure 3-19) To the  East Coast On February 20, 1979,

               1020    ,1016           ^        1016
       1C 7.0
                                    IOIG
                                                                                                  u>
                                                                                                  I
                                                                                                  U)
                                                                                                  en

-------
                               3-37
                   NDV    DEC   UHN   FEB   MRR   flPR    MRY    dUN
Figure 3-21.
Top: Daily  Sulfate,and Bottom: Sulfur  Dioxide
Concentrations (jug m~3) At the Experimental Lakes
Site (see Figure 3-3).

-------
                              3-38
                                                                   '.<-,
                            LONG POINT I
Figure 3-22.   Back Trajectories At 1000mb During February
               16-20, 1979.  (Crosses mark the  24-hour  time
               intervals)

-------
                               3-39



At Chalk River, Ontario  (Figure 3-22), the daily mean concentrations

of sulfate and sulfur dioxide were 13.8 and 100 jug m~"3, respectively,

This sulfur dioxide concentration was the highest value observed
                                         t
between November 1978 and June 1980.  Air parcel back trajectories

are shown in Figure 3-22 for Chalk River and Long Point and indicate

an air parcel origin south of the Great Lakes.  In southwestern

Ontario, daily mean S02 concentrations were 110 to 230 jug m~ .

Total suspended particulate (TSP) concentrations averaged 90 jug

m~"3, ranging from 30 to 130 ^g m~3.  About 30 percent of the

TSP consisted of sulfates.  Visibilities at southern Ontario

airports decreased from 20 km on the afternoon of the 19th (prior

to the incident) to 5-10 km on the 20th.

     A light mixed rain and snow event (accumulation of 2 mm)

collected at Long Point on Lake Erie on February 20 during the

episode had a pH of 3.8.  The concentrations of sulfate and

nitrate were 229 and 98 jug 1""^, respectively.  In contrast,

precipitation collected on February 23 and 25 had a pH of 4.1

and 5.6, respectively.

     Episodes of low pH values over a regional area have been

documented recently in the MAP3S/RAINE Oxidation and Scavenging

Characteristics of April Rains (OSCAR) Program.  The objectives

design, and synoptic overview of the OSCAR program has been

described recently by Hales (1981).  Results of an interesting

case study from the OSCAR program, when both the high and

intermediate density networks were activated, during the

-------
                               3-40


passage of a cold front with north-sourth orientation are shown

in Figure 23 and 24.  The pH of the first rain sample at each

station during April 22-24, 1981/ shows a fairly organized

pattern with the minimum pH values over the same general area

as for the annual mean (see Figure 3-1).  The change of pH

with time at four selected stations during April 22-24, 1981,

shows the pH values generally increased during the event at

all four stations.  The large changes in pH values within a

short period of time is an especially interesting feature of

these data to explain from more detailed analyses of the

chemistry and meteorology.

3. 5  Background SuIfur _in_ North Ame^rica_ East of the Rocky Mqun t a ins

     The models discussed in this report attempt to estimate S02

and sulfate concentrations and depositions from the spatial

distributions of emissions and wind and precipitiation fields.

For any given region area there are three basic sources of

atmospheric sulfur:

(1)  natural sulfur (biogenic or wind-blown dust) released within
     the modeling area;

(2)  anthropogenic sulfur released within the modeling area';

(3)  sulfur entering the modeling area from outside.  This sulfur
     may be of natural or anthropogenic origTn"!  If" all anthropo-
     genic sources within the modeling area were absent, the
     concentration of oxides of sulfur in air and precipitation
     would be by definition 'background' concentrations.  Background
     concentrations are caused by natural sources within and natural
     and anthropogenic sources outside the region.

     The modeling area of concern in this report is that portion of

North America east of the Rocky Mountains.  Within this area

-------
Figure 3-23,
pH of the First Rain Sample At Each Station In
the OSCAR Program Network During the Storm Event
of April 22-24, 1981.

-------
    Figure  3-24
i
Q.
Change of pH With Time A^Four Selected Stations
In the OSCAR Program Network During the Storm Event

of April 22-24, 1981.
Ji



1
5
4

1
5
4

2
5
4

C


^~

4
_—

^

2


)4









/



1


!
5

1
3
[j 	 '
>3

1
05
1 1 1 1 1 1 1 1 1 1 1 1 II
, i — i 	 , 	 MIIRRARH RRHHk MH
_j— ' APRIL 23-24
1 1 <
1 1 1 1 1 1 1 1 1 1 1 1 1 1 7
16 \7 18 19 20 21 22 23 24 01 02 03 04 05 06
UPTON, NY
ri APRIL 23-24

1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
14 15 16 \7 18 19 20 21 22 23 24 01 02 03 04
SCOTTDALE, RA _
APRIL 22-23
i r ' — ' ' ' f.
1 1 1 1 1 1 1 1 1 1 1 1 1 1 <
24 01 02 03 04 05 06 07 08 09 10 II 12 13 14
Ar-H . r-^~U
p 1 CHAMPAIGN, ILL
APRIL 22
1 1 1 1 1 1 1 1 1 1 1 1 1 III
06 07 08 09 10 II 12 13 14 5 16 \7 18 19 20
HOUR
                                                                                            10
                                                                                             I

-------
                               3.-43






anthropogenic sulfur emissions total 19.7 4- 6 Tg yr.""-1- while



natural emissions are 0.4 - 0.5 ± 0.1 Tg yr."1.  A breakdown of



these emissions for eastern and western Canada and United States



is given in Table 3-5.  The ratio of anthropogenic to natural



emission is about 24:1 in the east and about 16:1 in the west.



Therefore, within the modeling area, natural sources of sulfur



are unimportant compared to anthropogenic sources.



     Air entering the modeling area also contains some sulfur.



In any model, the concentration of that sulfur can be an important



boundary condition, assuming that the "hemispheric background level"



is significant.



     For eastern North America there are three major air-mass



sources; the Pacific Ocean, the Carribean and the Atlantic south



of 30° N and the Arctic.  Background sulfur concentrations in



airstreams entering the modeling area from each of these sources



will be discussed in turn.



     Pacific Air



     Pacific air crossing the west coast will have concentrations



of close to "hemispheric background levels."  As the air crosses



the coastal range, the orographic precipitation induced in an



air mass by the mountains acts as a scavenging curtain that



for the most part keeps sulfur originating in the Pacific basin



from reaching the lee of the Rockies.  Thus, background sulfur



concentrations just to the east of the Rockies are extremely low.



Measurements in northern Alberta during summer (Barrie and.Barrie

-------
                             3-44
      Table 3-5  A Breakdown of Natural  and  Antropogenic
                 Sulfur Emi_ss_i_qns  (T?g LS .jyr._~1 )	  	


East
West
Totals
CANADA
Natural
0.25!
0.25!
0.5
Anthropogenic
1.8
1.2
3.0





U.S.A.
Natural
0.40*
0.1 +
0.5
Anthropogenic
14**
2.7
16.7
!   EPS (1980)

*   Galloway and Whelpdale  (1980)

** EPRI/SURE II

+   Estimated assuming that natural  areal  emissions  between the
   Rocky Mountains and latitude 95°W amount  to  no more  than
   twice the terrestrial emissions  in  the eastern region.

-------
                               3-45

et al. (1981)  yield background sulphur concentrations of about
4 ug m~3 .  Further south downwind of California on the lee of
the Rockies, concentrations are about 2 jug m~^ .  It was pointed
out in Section 3.1 that one manifestation of the precipitation
scavenging over the coastal mountains is a pH value of 5.6 in
precipitation sampled downwind of the mountain.  At west coastal
locations, onshore trajectories are associated with precipita-
tion pH's of about 5.0 which is probably reasonably representa-
tive of hemispheric background values.
     Air from the Caribbean and Southern Atlantic
     In contrast to air masses from the Pacific, nothing prevents
the pollutants in Caribbean and southern Atlantic air-masses
from entering the modeling area.  In summer a common route for
this maritime tropical air-mass is up the Mississippi Valley and
then northeastward through southern Canada.  Background sulfur
concentrations in this air remain a controversial subject.
There is some evidence suggesting that the concentrations of
"background" sulfates and sulfur dioxide are significant
(Reisinger and Crawford, 1981).  Others contend that although
background concentrations are somewhat elevated relative to
the modified Pacific air masses they are insignificant compared
to the large anthropogenic emissions along the route commonly
travelled by this air-mass (Henry and Hidy, 1980).

-------
                               3-46





     Arctic Air



     Arctic air penetrates the modeling area along its northern



boundary most often during the period from November to April.  In



mid-winter continental arctic air can penetrate as far south as



Florida on rare occasions but, more commonly, as far south as the



central eastern states.



     The chemical content of continental arctic air has recently



been under intense study by many northern hemisphere (c.f.



Atmospheric Environment Arctic Issue, 1981) in connection with



a deterioration of visibility referred to as "arctic haze", which



is most pronounced in the winter season.



     It has been found that anthropogenic particulates consisting



mainly of sulfates, soot and organic carbon emitted at mid-



latitudes are the cause of the haze.  The atmospheric concentration



of these particulates as indicated by sulfate was an annual



cycle reaching a minimum value (close to zero) between June and



September and a maximum value of_2 ^ug/m^ between January and



March.  This is illustrated in Figure 3-23 from measurments made



at in the Canadian arctic by Barrie et' a_l. (1981).  In early



1980, weekly average sulfate concentrations ranged between 1.2



and 3. 5 jug m~^.



     Particulate pollution in the arctic air mass is distributed



over large areas rather homogeneously.  Little difference is



observed between sulfate concentrations at two sites in the



Canadian Arctic:  Mould Bay and Igloolik, about 600 km apart

-------
                               3-47


(Figure 3-25 and. 3-26).  Its annual cycle is consistent from

year-to-year.  Elevated wintertime concentrations of particulate

sulfur in the air mass owe their existence to the relative
                                                      i
inefficiency of atmospheric pollutant scavenging processes

during winter.  In arctic air masses, aerosols have a lifetime

of several weeks compared to. several days to a week at mid-lati-

tudes.  In standard textbooks the atmospheric residence time of

SC>2 and sulfate are commonly given as about 1 and 3 days respec-

tively, which in the light of these recent findings can be mis-

leading.  Furthermore, the longer -residence times in certain

regions and seasons help to explain why hemispheric background

sulfate, though less important than local or "long range transport"

effects, can still .be significant in producing acidic precipitation

episodes in remote areas, especially where the atmospheric con-

centrations of alkaline particles are small.  The predominant

sources of particulates in the the North American Arctic are, in

order of decreasing importance, Siberia, Europe and eastern

North America.

     There is evidence suggesting that the background concentrations

of sulfur oxides in a continental arctic air-mass that penetrates

the North America modeling area undergoes an annual cycle similar

to that observed in air nearer the pole.  Weekly average and

weekly minimum concentrations of sulfur dioxide and particulate

sulfate at ELA-Kenora are shown in Figure 3-27 for the period

November 1978 to December 1979.  The continental arctic air-mass

-------
                             3-48
Figure 3-25.   Locations of Sites In the Arctic Aerosol Sampling
               Network (M- Mould Bay, I- Igloolik and A- Alert)
               and of the American Sampling Site (B- Barrow).

-------
                                3-49
    3
E
O)
a.
CL
_J
D
to

CO
CO
LU
O
X
LU
    0 L
         I   I   I
                                    I   I   I    I
           CLOUD COVER
                                                     MOULD

                                                     IGLOOLIK
                        1979
                                      M~rA~T~M

                                      1980
                                                                 80
                                                      Qi
                                                      LU

                                                      8
                                                      U
                                                      0
                                                      O
                                                      g
                                                      u
Figure  3-26.
Weekly-Average  Excess-Sulfate  (non-sea  salt)
Concentrations  In the Atmosphere At  Mould Bay
and Igloolik  As Well As of Monthly Average Cloud
Cover In the  Arctic. Source: Huschke, 1969.

-------
                                     3-50
      rvi
     a .
     LTI
         2S
         20
          IS
          10
                   1
                       1
                           1
                               1
                                   1
                                        1
                                            1
                                                1
                                                    1
                                                        1 - 1 - r
                WEEKLY HVERHGE t MINIMUM RIR-CONCENTRHT IONS
                LOCUTION; ELR-KENORP,       PERIOD: NOVEMBER IBVB-DECEMBER 1979
                                                         	n
         2S
         20
         IS
     II X
      a
      LTl
               WEEKLY RVERBGE I MINIMUM RIR-CDNCENTRflTION5
               LOCUTION: ELS-KENORH       PERIOD: NOVEMBER ISVB-DECEMBER 1973
               i	r	1  _  i    i  -  i    i
             NDJFMRMJJR   5   0   N    D
Figure  3-27.     Diagram  of Weekly Average  and  Minimum SC>2  and


                   304  Concentrations At the  Experimental Lakes
                   Site (see Figure  3-3) Between  November 1978
                   and  December  1979.

-------
                               3-51





is usually present at this sampling location between November and



April.  SC>2 and sulfate background concentrations  (as  indicated



by the weekly minimum concentration) are a maximum between January



and March.  They average 1 and 0.8 ;ag m~3, respectively.  In



summer, they are below the detection limit (0.2 jjg m~3 S04=; 0.5



pg m   S02).



     The seasonal variation of precipitation pH in the polar air-



mass, as estimated by Barrie et al.  (1981) from observed particu-



late acidity, is shown in Figure 3-28.  The pH is expected to



vary annually from a CC>2 induced pH of  5.6 in summer to a late



winter/early spring pH of 5.0.  This cycle is confirmed by measure-



ments of the pH of snow in a Ellesmere  Island Ice Cap by Koerner



and Fisher  (1981) shown in Figure 3-29.



     There is some evidence that the background pH of snow in



northwestern Canada is depressed during winter.  From a snowpack



chemistry survey around an isolated thermal generating station in



northwestern Alberta, Barrie (1980) found the background pH of



snow that accumulated between December  1977 and January 1978 to



be 4.9 to 5.



3.6  Conclusions



     The Phase II Data Analysis review  included annual,



seasonal, and episode deposition monitoring results and their



interpretation using emission inventories, ambient concentra-



tion data, and trajectory calculations.  This review found



the highest precipitation acidity on an annual basis in the

-------
                              3-52
      6.0
      5.5
 a
a
uj    5.0
      4.5
          JFMAMJ   JASON  D1 J 'F 'M A  M  J
                         1979                            1980
Figure 3-28,
Acidity of Fresh Snowfall  At  Mould Bay Estimated
From Measured Sulfate  Concentrations, Hydrogen
Ion To Sulfate Ratios  In Aerosols and A Scavenging
Ratio of 2 x 105 (by volume).

-------
            Figure  3-29.
5.5--
as
ex
5.0-
Seasonal Fluctuations Ihe pH of  Snow  In  the
Agassiz Ice Cap Ellesmere Island  For  Each Year
Marked. Source: Koerner and Fisher, 1981.
                    TfT
                                                  62-
0
1 1 1 1 1 I 1
2
	 \ 	 1 	 1 	 1 	 r~
3
-r- i | i
4
i i
                                     Depth  of  Ice  (meters)

-------
                               3-54

Northern Hemisphere over eastern North America, Western
Europe, and Japan, with more alkaline precipitation over the
large continental areas of Western North America and Asia.
The cause of the slightly acidic precipitation along the west
coast of North America is not well understood, but could
be due to either anthropogenic sources which do exist or the
release of biologically-produced organic sulfur compounds
from the Pacific Ocean surface or both.  The zone of maximum
acidity in eastern North America stretches in a corridor
through Ohio and Pennsylvania into Southern Ontario.  Avail-
able concentration data at remote locations in eastern North
America generally indicate a summer sulfate maximum and a
winter S02 maximum with highly episodic behavior in pollutants
on a daily basis.  In addition, calculated dry depositions
of sulfur are found to be of comparable magnitude to wet
sulfur depositions especially close to source regions and in
the winter season.  Recent interpretations of both concentra-
tion and deposition monitoring data using trajectory calcula-
tions indicates that maritime tropical air masses from the U.S.
are the principal conveyors of elevated concentrations and
depositions in the extreme northeastern U.S. and southeastern
Canada, as opposed to continental polar air masses from Canada.
However, the source-receptor relationships based on back-trajec-
tory calculations and event data at single monitoring stations
are not always straight forward and contain some uncertainties.

-------
                               3-55





Some of the uncertainties are inherent such as trajectories in



precipitation systems and the non-linear chemistry, while others



can be reduced by analysis of data at more sites and for longer



periods.  Finally, the data analysis review concluded that



within the Phase III modeling area of North America, natural



sources of sulfur are unimportant compared to anthropogenic



sources.  In addition, the review found that background sulfur



concentrations enter the modeling area from the Pacific, the



Carribean and Atlantic Ocean south of 30°N, and the Arctic.



While the Pacific air masses are found to have concentrations



close to "hemispheric background levels" due to scavenging in



orographic precipitation, the Carribean and Arctic air masses



may contain somewhat elevated sulfur concentrations relative to



those in the modified Pacific air masses.  The predominant sources



of elevated sulfur concentrations in North American Arctic air



masses in the winter are thought to be, in order of decreasing



importance, Siberia, Europe, and Eastern North America.

-------
                          Chapter 4
   4.  THE ROLE OF MODELING IN THE DEVELOPMENT OF EMISSION
         CONTROL STRATEGIES AND AN AIR QUALITY AGREEMENT*
4.1  Introduction

     A long-range transport model  (LRT) is a description of

the physical processes involved in long-range transport in

the precise language of mathematics.  Relationships between

components of the physical system are replaced with logical

connections or mathematical equations.  Once a model has

been verified with monitored data it can be treated as a

"reasonable" analog of the real world.  Then we can investigate

causal relationships between the variables of the physical

system by routine manipulation of the equations.

     The system associated with long-range transport is

complex.  The model describing this system consists of a

large number of submodels corresponding to the components of

the system.  To keep the computing effort manageable, it is

usually necessary to keep the submodels of the LRT models as

simple as possible.  This means that the long-range transport

model may not incorporate all our understanding of the relevant

physical processes.

     In theory, we could construct a perfect LRT model if we

had all the information necesary to solve the mass, momentum
*See Addendum A for a personal critique of modeling by L.
Machta, U.S. Co-Chairman.  See Addendum B for a response to
Addendum A by H. L. Ferguson, Canadian Co-Chairman.

-------
                             4-2






and energy conservation equations.  In practice this is



impossible, and we have to be satisfied with using limited



information to estimate concentrations.  This means that it




is difficult to predict the concentration field corresponding



to a given set of observations used to characterize the state



of the LRT system.  However, we.can attempt to estimate



concentration field averages for the infinitely large ensemble



of possible concentration fields described by the set of



input observations such as velocity and precipitation.  This



introduces the concept of an ensemble.  Note that the ensemble



is defined by the input information used in the model to



estimate concentrations.  A "good" model should be able to



predict the ensemble average.  The ensemble is defined as the



average over a large number of individual model runs in which



only one or a few adjustable parameters are allowed to change.



It is important to reiterate that measurements of concentration



(or deposition) are expected to deviate from model predictions.



The concepts can be clarified through a simple analogy.



     Consider an office with five rooms which can be occupied



by five people.  There is no restriction on the number of



people who can occupy any one office at a given time and we



assume that the desire of any person to sit in any office



does not depend on the number of people already in the office.



In others words, the "concentration" of any one room can vary



from 0 person/room to 5 persons/room.   Given this information



can we predict the occupancy of a specific room at any instant?



The best we can do is to say that the "average" occupancy of

-------
                             4-3



the room is 1, a number which is obtained by dividing the

number of occupants by the number of rooms.  We know that this

ensemble average will differ from the room "concentration" at

any one instant.  To understand this a litle better let us

compute the number of ways we can arrange 5 people among 5

rooms with no restriction on the number of people in any given

room.  Using some basic combinational mathematics we can show

that this number is 126.  Let us denote the number of combi-

nations with a specified number of people x in a given room

by N(x). Then we can show that N(0) = 56, N(l) = 35, N(2) =

20, N(3) = 10, N(4) = 4, N(5) = 1.  These calculations indi-

cate that out of 126 possible combinations we are likely to
                                                              i
see 56 zeros, 35 ones, 20 twos, 10 threes and•4 fours and 1

five.  Note that the frequency of seeing a zero is the highest.

At any one instant the probability of seeing the ensemble mean

value of unity is only 28%.  The mean value over a period of

time during which the occupants have gone through a large

number of. rearrangements is
          C = 56x0 '+'35x1'+ 20x2 + 10x3 + 4x4 + 5x1       (4-1)
                                 126

            = 1.0

     How long do we have to wait before the average concentra-

tion approaches the ensemble mean C.  If the office is square

with sides of length L and u is the velocity with which the

occupants move around, the time taken for the rearrangement

-------
                             4"-4


of the offices is approximatley L/u.  This suggests that we

have to wait several of these time scales for the average

measured concentration to approach the predicated -ensemble

mean.  To make the discussion more quantitative let us calcu-

late the ensemble variance as follows

<(C -C)2> =  [56x(0-l)2 + 35x(l-l)2 + 20x(2-l)2 +10x(3-l)2

         . + 4x(4-l)2 +lx(5-l)2] 1 126
                 » l."33

Then, the expected deviation between measurement and prediction

(e2) as a function of averaging time:  .

          e2 = 1.33 L                                      (4-2)
                 U T

If we take L = 15m and u = 0.25ms"1, the time scale governing

rearrangements of the office is approximately 1 minute.  For e2

to become small, say 1/10,  the averaging T should be approxi-

mately 15 minutes.

We can make our model a little less simple by using an obser-

vation of the number of occupants in one room.  If j is the

observation, the prediction of the model is .

          C = (5 - j)/4                                    (4-3)


If j = 1, C = 1 and if j = 2, Cf = 0.75.  Note that the

ensemble mean prediction is a function of the way we define

an ensemble.  Specifically, the ensemble is defined by the

value of the input 'j1 to the simple model.  To consider the

implications of increasing the information content used in a

model consider the case when j = 1.  We now have 4 people who

-------
                             4-5


can move around in 4 rooms and the number of ways this can be

done is 35.  In terms of our previous notation we can write

N(0) =15, N(l) = 10, N(2) = 6, N(3) =3, N(4) = 1.

     Note that the probability of observing the ensemble mean

increases slightly to 29%.  The ensemble variance becomes


      <(C - C)2> = 15x1 + 10x0 + 6x1 + 3x4 + 1x9          (4-4)
                                35

                  = 1.20

     By using one observation, the variance of the prediction

decreases from 1.33 to 1.20.  Note that the variance is also

a function of the definition of the ensemble.

     If we refer to a model as complex when it uses more

observations, we see that the main difference between a simple

and a complex model lies in the expected deviation between

the model prediction and observation.  We can decrease this

deviation by increasing the number of observations.  However,

this will not hold true if the extra observations are in error

or if the complex model combines these observations in a

physically incorrect manner.

4.2  Uncertainties in Model Predictions

     The relationship between a model prediction Cpj_ and

observation Coj_ can be expressed as:

     Coi = Cpi + Ei                                       (4-5)

since the observations are expected to be normally distributed

about the ensemble mean, a "good" model would be characterized by

     E~i = 0                                               (4-6)
    <3T2 =    E.2  is small                                (4-7)
     t        i

-------
                             4-e
     Note that Ej_ is a measure of the accuracy of the model and
     is a measure of the precision of the model.  Then, the
uncertainty in the model prediction is expressed in terms of
the statistics of the residuals between model estimates and
observations used to verify the model.  The mean and variance
of E will not be useful if Coj_ is not normally distributed
about the ensemble mean.  Empirical evidence suggests that
concentrations are lognormally distributed.  This means that
residual statistics would be computed with log-transformed
concentrations.  In the next Section, we will show how these
uncertainty estimates can be used in the process of decision
making.
4 .3  Use of Model Results in Decision Making
     We have seen that model uncertainty can be expressed in
terms of residual statistics.  To illustrate their use we
will consider two exmaples of decision making, hydrology and
construction.
     Precipitation statistics are used routinely in the design
of dams and waterways (water control structures represent a
large investment activity on a continental, national or on state/
province scale).  Similarly, climatological wind statistics may
be used for designing skyscrapers or building bridges.
Often the decision to build a bridge has already been taken;
the question is how it should be constructed to be safe.  A
large safety factor is built in, even at significant extra

-------
                             4-7






cost, because the risks involved in not doing so are unaccept-



able.  These characteristics are food for thought in the



acid rain control context.  As in building bridges, where




the intangible of human life is at risk, the protection of



intangible environmental components is a major element to



consider.



     In the case of regional average frost free period statis-



tics and similar data, economically significant decisions are



made regarding the choice of major agricultural crop options.



Here again, such data (constituting the general planning model)



proscribe the envelope of feasibility or the broad decision



framework.  At a specific farm within the region, where more



detailed information is available or can be obtained, more



detailed plans can be developed.  For example, it will be



recognized that the annual frost-free period varies spatially



as well as from one year to the next, and that a farm situated



in a valley or frost hollow can expect to have a frost free



period shorter than the regional norm.  The planning strategy



to be used for that particular farm will represent a refine-



ment or subsequent level of planning within the general



regional "planning control parameters".  Decisions leading



to the optimum management of the farm are not necessarily



going to be made all at once.  The farmer may gather informa-



tion and experience over several years before achieving the
                  •


optimum.  This doesn't mean that he doesn't do anything for

-------
                             4-8.


that period of time.  In order to meet his particular objec-

tives he can implement a least-risk strategy based upon the
                                          v*
information available.

     The analogy between these climatic variables and current

concentration models  (and deposition transfer matrices) is

also valid in another respect, in that the annual precipitation

over a region follows a log normal type distribution.  Note

that one recognizes the possibility of an annual precipitation

value having a return period of 1000 years, but the building

of dams is not abandoned on the basis that construction for

flood control for the 1000-year storm is too expensive.

Instead, a more rational return period is selected and struc-

tures are installed on the basis that (a) flood control is

necessary and (b) there is a practical limitation to accep-

table' cost.

     In summary, LRT regional long-term models have similarities

to simple climatic models in the following respects:

     (1)  They do not incorporate everything we know about

          the physical problem (for reasons that can be

          readily explained);

     (2)  They should not be applied to problems for which

          they were not designed;

     (3)  They should not be dismissed or criticized on the

          basis that they don't explain phenomena on finer

          time and space scales than those for which they

          are intended;

-------
                             4-9





     (4)  They can be, and in the case of climate models



          have been, applied to major economic decisions in



          many sectors; and



     (5)  They have similar statistical variabilities in



          space and time which define significant (but



          accepted) error limits when applied to individual



          population elements.



     These examples illustrate the general nature of decision



making.  All decisions in the real world are made in the face



of uncertainty.  We are uncertain about the course of future



events.  Added to this is the uncertainty in our present



knowledge of the physical system we are interested in.  We



cannot discount LRT models just because they provide uncertain



concentration estimates.  We have shown that these predictions



can be useful if the decision maker explicitly accounts for



model uncertainty in making decisions.



4.4  Transfer Matrices



     A transfer matrix is an array of numbers linking the



sources to the receptors linear.  For example,



     Ci = Tij E                                            (4-8)



where Ej is the emission source j, C^ the concentration to



the receptor i and TJ_J is the transfer coefficient for



concentration.



     Here, i = 1, 2, 	M and j = 1, 2, 	N, and a



summation is carried out over repeated indices.  We can

-------
                             4-10





similarly formulate transfer matrices for deposition.



     Transfer matrices can be generated by LRT models.  Thus



the uncertainties in model outputs, described above, also



apply to the uncertainties in the elements of the transfer



matrix.



     Appendix 6 gives an example of how transfer matrices



could be used for control strategy evaluation if the transfer



coefficients had no uncertainty.  Since there is uncertainty,



the optimization must be based on probabilities.  This type



of optimization procedure will be examined during Phase III.

-------
                          Chapter 5


           5.  SUMMARY OF SELECTED MODELS AND THEIR
                  INTERCOMPARISON/EVALUATION

5.1    Summary of Model Profiles

5.1.1  AES -'LRT                                           ^

     The three-dimensional trajectory model uses objectively

analyzed wind fields and computes vertical motions at four

pressure levels:  1000, 850, 700, and 500 mb.  The analysis

procedure is essentially a three-dimensional scheme incorporat-

ing hydrostatic and height-wind balance routines (Rutherford,

1977), and producing gridded analyses of u and v wind components,

temperature, dew point and precipitation.. Input wind fields

are available every six hours and interpolation routines

are used to obtain winds at intermediate positions in time

and space.  The computations are performed on the standard

Canadian Meterorological Centre grid of 381 km at 60° N,

with the capability of operating on sub-grid scales down to

95 km.  Trajectory segment endpoints are determined each

time-step by assuming constant acceleration and using an

iterative scheme.  The motion of air parcels can be followed

backward (receptor mode) or forward (source mode) from

anywhere in North America.

     Trajectory paths are computed across grid cells of pollu-

tant emission, monthly mixing height and daily precipitation

amount.  Uniform vertical mixing is assumed to occur instan-

taneously up to the mixing height, and transformation and

removal processes are linearly parameterized (Olson, et al.,

1979.

-------
                            5.1-2

     The one-layer model parameterizes the physical and
chemical processes within a unit box extending vertically
from the ground to the inversion defining the mixing height.
                                       •*
Pollutant removal is parameterized by wet and dry deposition
and chemical transformation.  Pollutant input to each box is
provided from an annual North American SC>2 or NC>2 emissions
inventory on a 127 km grid  (Voldner and Shah, 1980).  Instan-
taneous mixing occurs throughout the box.
     The boxes follow trajectories that have been previously
computed and stored in 3-hour steps by the trajectory model.
At each step there is pollutant input from the inventory/
chemical transformation and surface deposition.  The combina-
tion of these processes results in a new concentration value
within the box.  The new concentration value is carried over
to the next point where the process is repeated.
     The SC>2 to S0^= chemical transformation rate is assumed
to be constant.  Dry deposition is parameterized in terms of
deposition velocities.  Wet deposition is parameterized by
scavenging ratios and by a gridded daily array of precipitation
amount.
     Trajectories from the AES trajectory model have been
compared to trajectories computed from various European and
American trajectory models  (Olson, et al., 1978).  Trajectories
computed at levels above the surface (i.e., 925 and 850 mb)
exhibit more agreement than surface trajectories.

-------
                            5.1-3





     A numerical analysis of the trajectory model has been



conducted and a report is presently being prepared (Walmsley,



et al., 1981).  An analytic non-divergent wind field was used



in the; intercomparison oK: model trajectory positions with analytic



solutions to the trajectory equation.  No serious deficiencies



were found in the formulation of the model.



     A short preliminary model evaluation (Olson, et al., 1979)



and a more complete model evaluation using EPRI-SURE data for



October 1977 (Voldner, et, al., 1980) have been reported.



     The AES-LRT model is periodically evaluated with additional



network data; a summary of the current evaluation statistics



for the Phase I model for January and July of 1978 and for the



entire year are given in Section 5.2.



5.1.2 ASTRAP



     The Advanced Statistical Trajectory Regional Air Pollution



(ASTRAP) model (Shannon, 1981) consists of three main subpro-



grams; vertical diffusion, horizontal dispersion, and calcula-



tion of surface concentrations and deposition.  Other programs



for data preparation or presentation of output are not con-



sidered a part of the basic model, since such programs are



usually tailored to a particular application.



     The vertical diffusion subprogram contains the dry



deposition and chemical transformation algorithms, as well



as a one-dimensional numerical solution of the standard dif-



fusion equation by the Gaussian Moment-Conservation technique

-------
                            5.1-4





(Shannon, 1979).  Dry deposition is parameterized by deposition



velocities; the major difference between dry deposition in



ASTRAP and that in most other LRTAP models are that deposition



velocities in ASTRAP vary with time of day and season, and



that average SC>2 and 804 deposition velocities are almost



equal in magnitude.



     The chemical transformation rate also has diurnal and



seasonal variations.  There is an increased transformation



rate during initial dispersion (first three hours) for emis-



sions from the lower layer, in order to simulate the effects



of increased catalytic transformations in more polluted urban



areas.



     Eddy diffusivity profiles which simulate the cycle of



nocturnal surface inversion formation, deepening, and erosion



from below, are specified by hour and season.



     A diurnal variation in the emission rate is also included



in the vertical diffusion subprogram.  The variation is a maxi-



mum in the surface layer, and decreases with height until



the variation about the average emission rate becomes zero



in the sixth layer (600-800 m).  The variation of emission



rate is an arbitrary estimation of the effect of diurnal



variations of heating and cooling loads, working hours, and



the like.

-------
                            5.1-5





     The number and thickness of the layers in ASTRAP is



optional.  Since the eddy diffusitivity profiles and emission



inventories are specified for each layer/ they must be re-

                                                                i

estimated or relocated each time the layer definition is changed,



     Vertical profiles of one-dimensional concentrations of



SC>2, primary 804, and secondary 804 (i.e./ that produced by



atmospheric transformation of SC>2) are calculated for norma-



lized emissions within each layer in turn.  The advantage of



calculation in this structure is that scenarios of varying



fuel type/ or source region, or scenarios about different



stack heights, can be examined later without recalculation



of the profiles.                                    .



     The horizontal dispersion subprogram utilizes the concept



that long-term diffusion in the horizontal is determined by
                        \


the distribution of the plume centerlines rather than by



small-scale diffusion about the centerlines.  The statistics



of endpoint locations are calculated and stored for six-hour



increments.



     Wet removal is simulated by first advecting the tracer



for a six-hour time step and then checking the new location



to see whether precipitation has occured during the six-hour



period.  If so, a fraction, F, of the sulfur mass represented



by the tracer is deposited as a function of the half power



of the precipitation amount (Hicks and Shannon, 1979).  The



tracer portions deposited are stored by plume age for each

-------
                             5.1-6




 source,  and  statistics  similar  to  those  for dry  tracers  are



 generated.   The  fractional  "dry" tracer  remaining  after  a


 precipitation  event  is  (1-F)  times  the fraction  at the



 beginning  of the event.  The so-called dry tracer  is  the one


 not  subject  to wet removal.



     The statistics  generated by the horizontal  dispersion



 program, for a grid  of  virtual  sources and for each increment


 of plume age,  are the mean  position and  spread of  the end


 point ensembles  and  the number  of  equivalent dry tracers con-



 tributing  to the statistics,  for both dry tracers  and wet-


 deposited  tracers.   Note that these statistics are independent


 of sulfur  species, because  the  wet  removal applies to bulk
                                   »


 sulfur.


     Statistics  from the main subprograms are combined with


 an emission  inventory,  including primary sulfate emission



 factors, to  produce  output  in a non-normalized form.  The sub-


 program  combines the horizontal distribution statistics  of dry


 tracers  with the normalized  one-dimensional surface concentra-


 tions and  dry  deposition increments; combines the  horizontal


 distribution statistics of  wet  tracers with the  normalized


.one-dimensional  budgets; and  sums  the resulting  concentrations


 and  deposition from  each source for a regularly  spaced receptor


 grid, or for a list  of  receptor locations.

-------
                            5.1-7



     Basic products of ASTRAP simulations are regional long-


term average (monthly or longer) fields of S02 and 804 con-


centrations, and cumulative wet or dry deposition of total


sulfur.  Dry deposition can he subcategorized as dry deposi-


tion of SC>2 and 804; however, the wet removal parametrization


is for bulk sulfur and thus speciation is artificial.  By


integration of the deposition fields for specified subregions


of the grid (such as the eastern U.S. or eastern Canada)'


sulfur budgets are also obtained routinely.  In general, ASTRAP


should only be applied in larger meso- to regional-scales and


monthly or seasonal time scales.


     Adjunct subprograms of ASTRAP objectively analyze wind
                 »

and precipitation fields, grid emissions horizontally and


vertically, and contour simulated concentration and deposi-


tion fields on a background map of eastern North America.


     The wind fields for ASTRAP are computed by first


calculating the mean wind between the surface and 1 km for


each rawinsonde observation, and then performing a form of


inverse distance-squared objective analysis at regularly


spaced grid points.  The objective analysis scheme obviously


cannot improve upon the 12 hr and 200 km - 400 km resolution


of the raw data.


     Hourly precipitation observations are summed for six


hours, combined with 6-hour precipitation observations, and


analyzed for a grid spacing of about 70 km.  If no observation


site is inside a grid cell, the nearest observation is used.

-------
                             5.1-8




      Emission  data  can  be  either a  list  of  point  sources,


 with  appropriate  location  and  stack data, or a  grid  of  virtual


 sources,  sorted by  effective emission  layer in  either case.


 5.1.3  ENAMAP


      In  the mid-1970's,  SRI, International  developed the


 trajectory-type European Regional Model  of  Air  Pollution


 (EURMAP)  for the  Federal Environmental Office of  the Federal


 Republic  of Germany (Johnson et al.,  1978).   In the  late


 1970's,  the U.S.  Environmental Protection Agency  contracted


 SRI,  International  to adapt  and apply  the EURMAP  model  to


 eastern  North  America.   The  adapted model,  ENAMAP (Eastern


 North  American Model of  Air  Pollution),  is  capable of calcu-
k

 lating long-term  SC>2 and 864 concentration  and  dry and  wet


 deposition patterns and  regional and  international exchanges


 resulting from the  emissions of S02 and  804.  It  should be


 noted  that another  version of  the model, ENAMAP-2, is currently


 being  produced.   This second version  will upgrade the para-


 meterizations  of  vertical  mixing, dry  deposition, and trans-


 formation and  will  include nitrogen chemistry by  the end of


 1982.


     Basically, the ENAMAP model can  be  classified as a mass-


 conserving Eulerian-puff model.  Chemical processes  within each


 puff  are  parameterizated as  they move  across the  modelling


 domain.   At 3-hour  intervals,  the concentrations  and deposi-


 tions  are calculated from  each puff and  apportioned  to  the


 grid  cells on  the basis  of the portion of the puff within


 each  of  the cells.

-------
                            5.1-9





     Discrete puffs of S02 and 804 are released every 12 hours



from 80 km by 80 km emission grid cells.  The mass of pollu-



tant in each puff is determined by dividing the annual emis-



sions by 730, the number of 12-hour periods in a year.  These



puffs are tracked in 3-hour time steps until either they move



outside the model domain or their concentration drops to an



insignificant level.



     The individual puffs are transported in a vertically-



averaged and horizontally-interpolated wind field.  These



fields are updated every three hours.



     Upon release, each puff is assumed to immediately diffuse



vertically to yield a uniform concentration in the layer be-



tween the surface and the mixing, height.  In the model, the



mixing height varies seasonally, between 1150 and 1450 metres.



The constant transformation rate of S02 to S04= (1%/hr) in the



ENAMAP model was chosen after a review of field, laboratory, and



theoretical studies.  The SC>2 and 804 dry deposition rates are



treated as constant throughout the simulation period.  The dry



deposition rates, representing the daily average, are based



on reviews of fields, laboratory, and theoretical studies and



on an evaluation study conducted with the EURMAP model.



     The wet deposition calculations are based upon hourly



precipitation rates and the duration of puff exposure to that



precipitation.  The model does not distinguish between rain



and snow scavenging.  Every three hours, the model determines

-------
                            5.1-10

via preprocessed, objectively-analyzed, three-hourly precipi-
tation fields the precipitation intensity in the vicinity of
the puff.  Within each three-hourly simulation period, an
average hourly precipitation rate is estimated.
     The monthly SC>2 and 804 deposition patterns are obtained
via summation of the contributions of each puff within each
receptor grid cell.  The monthly-mean SC>2 and 804 concentra-
tion patterns are obtained in a similar manner.
     Transport wind fields were generated at 3-hourly intervals
using the receptor grid network of 70-km spacing and the 12-
hourly surface and upper-air wind data available at approxi-
mately 60 sites in the United States.  At each data site,
average u^ and \r components of the transport wind in the sur-
face layer were calculated.  The transport winds were then
generated from these layer-averaged wind components by a
distance-weighted interpolation scheme using a weighting factor,
Precipitation fields were generated at 3-hourly intervals
for the same 70 km grid network.
     Since the ENAMAP model was designed to consider primary
sulfate emissions, both S02 and 504 emissions were gridded
separately on an 80 by km UTM grid network.  The annual
emission grids were generated by simply adding the annual
emission rates from all the point sources within each grid
square and the county-wide area sources of those counties
whose geographical center was locked in the grid square.

-------
                            5.1-11






     No attempt was made to consider natural emissions of S02



or 804.  Emissions crossing the western boundary and entering




the modelling domain also were not considered by the model.



5.1.4 OME-LRT



     The OME-LRT model is statistical in that the physcial



processes of transport are expressed in terms of statistical



parameters.  The basic premise of this class of models is



that long-term concentrations are insensitive to short-term




fluctuations in meteorology.  It is assumed that concentra-



tions averaged over periods of the order of a year reflect



"mean" patterns of large scale meteorology.  This allows one



to take a simple approach to the modelling of long-range



transport.



     The model is based on the idea of classifying pollutant



particles as "wet" or "dry".  Wet particles exist during



precipitation and dry particles during dry periods.  Over a



long term, each travel time from a source is associated with




a particular mass of dry particles and a particular mass of



wet particles.  Using this concept one can formulate differen-



tial equations for the evolution of these particles as func-



tions of travel time from the source.  It is assumed that the



average rate of "conversion" from wet to dry particles is



inversely proportional to the average length of wet periods



in a Lagrangian sense.  A similar assumption can be made



regarding "conversion" of dry particules to wet particles.

-------
                             5.1-12

Further,  it  is assumed  that  the  scavenging  coefficients  do
not vary  with travel  time, but are  different  for  wet  and dry
periods.
     The  transformation of SC>2 to 804  is  a  complex  process
that depends on  a  number  of  physical variables  such as  solar
intensity and ambient ozone  concentration (see  Wilson and
'Gillani,  1978).  For  long-term modelling  we assume  that  the
conversion rate  is  1%/hr, a  value which is  an "average"  of
field measurements  made during dry  periods.
     The dry deposition velocities  chosen,  1.0  cm s"1 for S02
and 0.05 cm  s"1  for S04,  are based  upon previous  modeling
efforts and  some field  measurements.   However,  characteriza-
tion of dry deposition  remains a contentious matter among field
experimentalists.
     A table in  the complete OME-LRT model  profile  presents
all the model parameters  used in the simulations.
     The  long-term  concentration C(t,  t)  at point r at  time  t
can be written as  (Lamb,  1980)
     _              t
     C(r_, t) = Q  r  p(r, t|rs_, t1 )dt'                  (5-1)
                  — oo
where Q is the emission rate of  the source  located  at rs and
p(r> fcl£s' t') is  the probability density that  a  particle
released at  rs at  time  t1 will be found at  r at time  t.   We
             -o3                             ^*^
assume that  scavenging  and dispersion  are independent.

-------
                             5.1-13

     The dispersion function will depend upon  large  scale
wind patterns.  The parameters of the distribution x" (t),
y (tr) are the coordinates of the mean position of the
particles position after travel time 'tTfrom  the  release
point.  The dispersion parameters C)x~ (t) and CJy('C) correspond
to standard deviations of particles about  (3c,y)  after  travel
time "C from the source.  These parameters  can  be determined
from trajectory statistics as suggested by Bolin and Persson
(1975).
     The wet and dry deposition of sulfur  depends on the
vertical distribution of the pollutant as  well as the  turbu-
lence in the diurnally varying planetary boundary layer.
The distribution of SC>2 and 864 was taken  to be  uniform
in the vertical through the depth of a constant  mixed  layer.
We should point out that this limits the resolution  of the
model to distances of the order of 100 km  from major sources.
     The large scale horizontal distribution of  pollutants
is determined by the parameters x/y/ <3x and  <3y.  For the
coordinates of the mean motion of large scale  eddies we
assume that
     x" = u tr
                                                      ( 5-2)
     y = 0
where u is the mean velocity of synoptic eddies  and  is the
travel time from the source.
     The analysis of trajectories by Slinn et  al (1979) and
Bolin and Persson (1979) suggests that Ox"  and  dy" can be
expressed as

-------
                            5.1-14
     Oy =
                                                    (5-3)
     On the basis of statistical dispersion theory it is

reasonable to assume that  u and  v are the standard deviations

of the horizontal velocity fluctuations of synoptic turbulence.

These statistics can be derived by sampling 850 mb winds over

periods of the order of years.  Tennekes (1977) suggests

the following values for the large-scale velocities.

     u = 10 ms'1,  u = 10 ms"1,  v = 6 ms"1

     The emission data for the model was collected from

several sources.  In the northeastern sector of the U.S.,

it was compiled from the EPA point source inventories

(Benkowitz; 1979) and the GCA  (consulting company cotracted

by EPRI) major point and area source record.  The Canadian

emission points except in Ontario were taken from the AES

preliminary point and area source inventory (Voldner et at.,

1980).  The Ontario points were extracted from the Ontario

Ministry of the Environment sulfur dioxide emissions inventory.

     All large (> 100 kTonnes/yr S02) point sources listed

in the above mentioned emissiond data and 95% of the major
                            V
(>10 kTonnes/yr S02) point sources were incorporated into

the model's inventory (usually grouped) to form effective

point sources located at the emissions-weighted geometric

means of the coordinates of the contributing points.

Approximately 60% of all area emissions and 72% of all minor

-------
                            5.1-15

point emisions were incorporated into the inventory by adding
minor point sources and area sources located near (  50 km)
major points to that point or combining small sources concen-
trated in large urban centers to form effective point sources.
     The model sensitivity to uncertainty in the value of the
input parameters was tested by independently varying each
parameter within the range of values cited in the recent
literature.  For this test an idealized source-receptor
to the input parameters as a function of the source-receptor
orientation.
     The interested reader is referred to the OME-LRT model
profile for the completed discussion.
5.1.5  RCDM-2
     A simple approach which gives the same basic results
as the more computationally involved methods has been proposed
recently by Fay and Rosenzweig (1980).  These authors have
assumed that the longer period sulfur dioxide and sulfate
concentrations from a point source can be described by the
steady state diffusion equation in which the horizontal eddy
diffusivity and conversion and removal rates are uniform in
space.  Analytical solutions to the diffusion equations for
sulfur dioxide and sulfate concentrations are found under
these simplifying assumptions.
     The sulfate predictions from the steady state model are
also in general agreement with those from the ASTRAP model.
RCDM-2 preserves the basic features that produce essentially

-------
                             5.1-16


the same mean transport field that one gets from a large

number of trajectories but eliminates most of the detailed

fluctuations.  Seasonal and  annual resultant wind vectors

at the upper air stations are used.

     The two-dimensional steady-state advection-dif fusion

equation with removal is:

     u ]*C    +   v 3C  = Df }2C   +   "^2CN\ -        (5-4)
where u and v are the mean wind velocities in the x and y

directions/ D^ is the diffusivity and f is the removal  (wet

plus dry) time constant.  This equation can be solved for

a steady point source at the origin having an emission rate

Q, with a boundary condition of. zero concentration at infinity;

C =   Q      expf w x*\ K frfl   + fj2_Y\i/2 V       (5-5)
                ^2^7  '   \  [_Dh-C   \2Dh/ J    J
where the x* axis is aligned in the direction of the mean wind.

velocity, w, Ko is the modified Bessel function of the zeroth

order, r is the radial distance from source to receptor, and

h is the mixing height.  The completed deriviation of this

equation appears in RCDM profile report.

    For a gridded emission inventory, the model assumes that

emissions are concentrated at the center of each square.

Because of certain model assumptions, there is no reason

to expect predicted concentrations in home grid squares to

be realistic, but they still must be finite, and still should

be consistent with model behavior in adjacent grid squares.

-------
                            5.1-17

For these reasons, the recent TRI work (Benkeley and Mills,
1980) computed home grid square concentrations as the area
integral (using Simpson's rule) along the mean wind axis in
the home grid square from its center to its edge.  This
turns out to be equivalent to solving the equation at a
distance of 23.5 km. from the origin and this distance is
then used for all home square calculations.
      The analytical solutions for the horizontal distribution
of primary and secondary pollutants from a steady point source
at the origin having an emission rate Q are:
C-, = _ Q_  exp/ w x'\   K \ r [~1   + /w  \*\ V2 Y    (5_6)
                   '
and
C2 =  fiQ     exp(_w_ xA  K0(Yr) - KQ(^r)
                   h   j    oc.2  -  ff2
                                                      (5-7)
respectively where
     o,2 =  /_!_ +  _wL"\                             (5-8)
           V^°n    ^l  1
     x2 =  / _i _ +  -Sii-^                            (5"9^
           V-TTDh    4D^  )
    Here the x' axis is aligned in the direction of the mean
wind velocity, w, Ko is the modified Bessel function of
zeroth order, r is the radial distance from source to receptor,
h is the height of the mixing layer, D is the horizontal
diffusivity, and p is the mass ratio of secondary pollutant
formed per mass of primary pollutant.  The rate constant

-------
                            5.1-18







takes into account all forms of depletion of the primary



pollutant.  The rate constant for loss of secondary pollutant is
                "  = ( T: g"   +   r')                    (5-10)



where the rate constants for wet and dry deposition of the



secondary pollutant are ""£~q   and"C^.~ , respectively.  Contri-



butions from multiple sources of primary pollutants can be



superimposed since the boundary condition at infinity has been



specified to have an ambient concentration of zero.



     The Modeling Subgroup Report (section 2.5) and the RCDM



Model Profile contain more description of the RCDM model inputs,



sensitivity, and evaluation results.



5.1.6 CAPITA Monte Carlo



     The Monte Carlo approach to simulation of physical and



chemical atmospheric processes has the following key characteris-



tics:  (1) each process (transport, transformation, deposition)



is simulated directly as a discrete event; (2) mass conservation



is maintained by counting each pollutant quantum as it moves



and changes chemical form, rather than via differential



equations; (3) direct Monte Carlo simulation is Lagrangian



since trajectories of individual quanta are followed, but



Eulerian in that in the limit of an infinite number of quanta



and infinitesimal timestep a stochastic solution of a two



dimensional diffusion equation with non-steady state



inhomogeneous flow is obtained.  Its Lagrangian features are



most useful for simulation of kinetics; the Eulerian aspects



benefit the transport simulation.

-------
                            5.1-19
     The CAPITA Monte Carlo model represents an attempt to
provide a conceptually and computationally simple approach
for daily simulation of air pollutant concentration and
deposition on the regional scale.  The model is continually
being modified to test alternative hypotheses; the discussion
below describes a currently operational version which is not
expected- to change dramatically in the near future.
     Emissions are represented by virtual point sources/
representing arbitrary mixtures of actual point and area
sources.  The emission grid spacing is nominally 190 km,
except in the northeastern U.S. and southeast Canada, where
the grid spacing is 95 km.  The emission grid, on a polar
stereographic projection true at 60°N, does not include
natural sources nor western North America sources.
     The advection of pollutants is facilitated using surface
data.  The rationale for the use of surface winds instead of
the more commonly used upper air winds is given by Patterson,
et al., 1981.
     The data for all reporting sites for each parameter were
interpolated onto a square grid via a weighting function.
Each gridpoint weighs about 3 to 6 of the nearest National
Weaher Service sites.  A further smoothing is applied to the
grid.  The wind vector components northward and eastward were
gridded and smoothed separately.
     The general features of the wind vector field from upper
air and surface wind data are similar.  However, it is evident

-------
                            5.1-20

that the higher station density network of surface data yields
more structure in the wind field.
     Surface wind speed is always less than the wind
speed for the bulk of the planetary boundary layer.  A
calibration factor is required therefore by which the surface
wind speed is increased to match the mean "mixed layer" speed.
This factor is obtained by taking the ratio of the properly
averaged upper air wind speed to the surface wind speed for
every grid point.  The seasonal variation of the scale height
over which the upper air winds are averaged were taken from a
study by the Atmospheric Environment Service of Canada
(Portelli, 1977).
     The concept of "the trajectory" of an air parcel during
multiday transport has no physical realization.  After repeated .
subjection to diurnal planetary boundary layer dynamics of
turbulent diffusion, oscillations in mixing height, and
vertical wind shear, a puff of emitted material may be dispersed
beyond recognition into part of a large scale "background" airmass.
     For computational simplicity, the "diffusion" is currently
simulated by an effective diffusion coefficient, K, such that
a displacement of radius  /2KAt is imposed with randomly
chosen direction following the advection step.  In the limit
of large numbers of quanta, the Monte Carlo approach is both
Lagrangian and Eulerian.  In the model simulation, only one
trajectory is chosen, so that the single quantum emitted at a

-------
                            5.1-21
source in the mid-West might be exported to the Great Lakes
within 3-4 days or remain in the Rocky Mountain states after
10 days, depending on the sequence of random diffusion steps
it experiences.
     The general features of the simulated spatial distribution
are preserved without distinct plumes if a value of K = 105m2s~l
is assumed, but K = I06m2s~l causes an overly uniform simulated
spatial distribution.  We have adopted a value K = 4 x lO^m^s"-'-.
     Just as for the random walk nature of diffusion, the
CAPITA Monte Carlo approach treats the kinetics of chemical
transformation, dry deposition and wet removal as stochastic
events.  The model requires specification of transition
probabilities.  Over each timestep, all possible paths (S02
conversion to 804, etc.) are assigned a probability of
occurrence. Equivalent simulation of the ensemble mean is
achieved by algebraic allocation of mass among the various
forms, treating the transition probabilities as expected
fractions of conversion and removal.
     Currently, the model assumes that S02 emissions contain
1% primary S04=, and lumps wet and dry deposition together.
The values used are constant diurnally.
     For the Phase II effort,  the SC>2 deposition was presumed
to be 95% dry deposition and 5% wet.  Sulfate was allocated
such that 20% was dry and 80% of 804 deposition was via
"precipitation".

-------
                            5.1-22






     Simulation .of wet removal will be included in Phase III



of the MOI activity.  The formulation requires knowledge of



the probability of the pollutant experiencing precipitation



within a timestep.  From the spatially dense hourly precipitation



data, the number of reports of precipitation divided by number




of reporting sites yields the likelihood of rain.  Removal



percentage for both SC>2 and 804, given that the pollutant is in



a precipitating airmass, then multiplies the probability of



precipitation to yield the desired transition probabilities



of wet removal.



 ^  The CAPITA Monte Carlo model has been compared to sulfur



concentration and deposition data for both episode studies



(Patterson et al., 1981) and seasonal averages.  Earlier



studies indicated that the transport so dominates that the




simulation is somewhat insensitive to both source specification



and kinetics for episode studies.



     The model evaluation and sensitivity studies, as adopted



by the Modeling Subgroup, are presently being performed for



the CAPITA model.



5.1.7  MEP-TRANS



     A regional trajectory based model, TRANS (Transport of



Regional Anthropogenic Nitrogen and Sulfur) was developed to



satisfy both requirements of resolving the patterns of single



plumes near the sources and the effects of large scale atmos-



pheric motions in transporting the plumes over distances of



one thousand kilometers or more.

-------
                           5. 1-23






     The horizontal wind field is obtained from the sea-level



pressure field distribution through the so-called geostrophic



approximation with some adjustment being made in both speed




and direction in order to be representative of the 300 to 500



metre level above the surface.



     The observational data consists of sea-level pressures



obtained at 6-hourly intervals on the CMC 381 km grid.



     A method due to Sykes and Hatton (1976) is used in which



orthogonal polynomials in the two space co-ordinate x and y,




and in time co-ordinate t, are fitted to the data.  The space



polynomials are defined on the network of observing points by



recurrence relations.  The series of six-hourly observations



for the day allow the determination of pressure at any time



between observations.  The spatial pressure field is fitted



within one millbar error corresponding to a wind field error



of the order of one m/s.



     Chung (1977) has derived a method by which a wind field



representative of the atmospheric motions at a height of



300 to 500 meters above the ground may be obtained from the



geostrophic wind.  In this approach used in TRANS, both



speed and direction adjustments are obtained by means of a



regression equation in terms of the geostrophic wind speed



and the 3-hour pressure tendency.



     The trajectory integration is done by a method due to



Peterssen (1956), in which an iterative procedure is used to



determine the new position after travel time At.

-------
                            5.1-24






     The integration can be carried forward in time to deter-



mine the motion of the air parcel from an arbitrary location,



and can be carried backward in time to determine the history



of air parcels arriving at an arbitrary location.



     The accuracy of the integration procedure.was determined



from tests carried out over 96 hours of travel time in both



forward and backward trajectory modes with a 3-hour timestep.



It showed a maximum difference in spatial position of 1 km,



indicating a negligible source of error in comparison with



the uncertainties in the pressure fields.



     The pressure field uncertainties are due to fitting errors



and errors in the original data.  Only a small fraction of the



observational data error (about 0.5 mb) is being transmitted to



the interpolated values for the missing data.  The fitting



procedure introduces an error in the pressure field but tends



to eliminate random noise inherent in the original data.  Frontal



features are usually not resolved by the pressure data, leading



to error in the determination of the trajectories near fronts.



Typically, for randomly oriented velocity errors, trajectory




positional error may be of the order of 50 km after 1 day's



travel.  Where a sharp gradient in velocity exists, as in a



strong depression, the position of an air parcel after 24



hours travel can be several hundred kilometers from the true



position.



     The pollutant material contained in the air parcel



centered on the plume trajectory is assumed to disperse with

-------
                            5.1-25

time, so that the horizontal distribution about the centerline
can be approximated by a normal (Gaussian) distribution.  The
vertical distribution is assumed to be uniform.
     The mixing height is allowed to undergo a standard
diurnal cycle ranging from 0.2 to 2.0 times the seasonal mean
value in four discrete vertical steps.  The pollutant material
in each of the four layers is transported by the same wind field
and participates in the wet deposition.  Only the pollutant
material below the mixing height is assumed to be subject to
dry deposition.
     The deposition of material from the plume onto the ground
is modeled by means of the deposition velocity concept.
     Within the model/ the deposition velocity is dependent
on time but is not explic.ity made variable in the spatial
coordinates.
     Washout or rainout is parameterized by means of a bulk
washout parameter.
     The washout coefficient is proportional to precipitation
rate and is specified for each plume segment separately.
Precipitation rate for a given segment is determined from 3-
hourly precipitation intensities over the network of observing
stations.  Thus the washout of material is governed by the
precipitation encountered by each plume segment in its travel
path as determined by historical record.
     For the case of SC>2 washout, the approach suggested by
Barrie (1981) has been adopted.  This approach allows the

-------
                            5.1-26






temperature and pH dependence of the SC>2 washout ratio to be



explicitly incorporated.  The chemical transformation of 302



to 304 in the plume is modeled as a first order reaction.



The implementation of the transformation in the model is through



modification of the source terms after each time step.



     It is found that for normal conditions, the decay time



of SO2 is about 24 hours and total sulfur, 45 hours.  Similarly,



noontime emission produces a S02 decay of"time of about 36



hours and total sulfur decay of time 72 hours.  Similar



computations for NC>2 and total nitrogen, give 16 hours and 72



hours respectively, for standard conditions. -



     A high transformation rate k of S02 to 804 of 0.025 per



hour generates and maintains 10 times as much 304 in the



plume as does the low rate of 0.001 per hour.  The total



sulfur in the plume at long travel time is twice as high for



the high transformation rate, due to the lower sulfate



deposition velocity.



     It is found that the effect of the deposition velocity of



304 on sulfate levels indicates that only a 50% variation is



evident after 24 hours for deposition velocities in the range



of 0.1 to 2.0 cm/s.



     An effective mixing height of 1200 metres leaves twice



as much SC>2 and nearly 50% more sulfate in the plume after 24



hours as does the 500 m mixing height.

-------
                            5.1-27





     Moderate rain (5 mm hr~-M produces a very rapid removal



by plume washout/ the S02 decay time decreases to 9 hours



with what little sulfate forming initially being washed out



very rapidly.



     The nitrogen behavior was found to be similar to the



sulfur behavior described above.



     The emissions data for SC>2 and NOX used in the modeling



study were the 127 km griddded data currently being used by



AES in their regional transport evaluations.



     A total of 70 trajectory origins were selected to coincide



with grids having a large -emission rate.  For each trajectory



origin, 5-day trajectories at 3-hourly time steps were



generated for the full year.  The trajectories are used in



conjunction with the 3-hourly precipitation data to define



the precipitation intensity for each segment of the plume.



     Concentrations and loadings were calculated at 75 receptor



points.  The receptor locations were chosen to coincide with



the 11 sensitive receptor areas chosen by the working group



and a number of monitoring stations in the APN, CANSAP, MAP3S,



NASP, EPRI, NADP and Ontario Hydro networks to facilitate



subsequent comparison with data.  The values of the parameters



for the 1978 comparison where selected based on the sensitivity



tests.



     The MEP model annual S02 concentration ranged from



1 ug/m^ in Northern Ontario and Northern Quebec to 30 ug/m^

-------
                            5..1-2 8





over a wide region of the Eastern U.S./ with an isolated peak



of 50 ug/m3 in the Ohio Valley region.  The annual average



SO^ concentrations ranged from 1 ug/nr in Northern Ontario and



Quebec to 20 ug/m3 in Eastern U.S. with a peak of 25 ug/m3 in



the Ohio Valley region.  The N02 concentrations range from



0.5 to 20 ug/m3 with two regions of higher concentration



in the Detroit and New York areas.  Predicted NO^ ranged



from 1 ug/m3 in Northern Ontario to 20 ug/m3 in the New



York area.



     Maximum dry deposition of S02 of the order of 25 kg.S.ha."1



yr."1 occurred in association with the high concentration



areas.  The 804 dry deposition was approximately 10% of the



S02 deposition.  Dry deposition of nitrogen was up to 5 kg.N.ha



yr.   for N02 and 1 kg.N.ha.^yr.   for NOj".



     The predicted wet deposition of S02 ranged from 0.5



kg.S.ha.~lyr.-1 in Northern Ontario and Quebec to 15 kg.S.ha."1



yr."1 in Ohio with an isolated peak of 20 kg.S.ha.^yr.~1.  The



wet deposition of SO^ reached 10 kg.S.ha.^yr.   in the Ohio



region.  Wet deposition of nitrogen ranged from 0.1 to 5 kg.N.ha."1



yr."1, being approximately equal for the two species.



     Wet deposition of sulfur ranged from 1 kg.S .ha.^yr .-1 in



Northern Ontario and Quebec to a peak of 30 kg.S.ha.'iyr."1 in



Ohio.  Wet depositions of nitrogen ranged .from 0.5 to 10



kg.N.ha.^yr."1 in the lower Great Lakes regions.

-------
                            5.1-29



     Total (wet plus dry) sulfur deposition showed a large


area of excess of 30 kg.S.ha.~lyr."-^ centered in the


Ohio Valley and extending into Southern Ontario.  Total


nitrogen depositions expressed as N are approximately one-


third of the sulfur depositions on a regional basis.


5.1.8  UMACID


     The Atmospheric Contribution to Inter-regional Deposition


(ACID) model is a receptor-oriented model designed to estimate
           V

the contributions of upwind sources to measured pollutants at a


given receptor.  The model incorporates the horizontal diffusion


due to vertical wind velocity shear, chemical transformation,


spatially and temporally varying dry deposition and "events"


precipitation scavenging.


     The model is intended for use at rural locations where


local sources are negligible.


     The use of a mixed layer trajectory model assumes that


the material being traced is moving with the mean motion of


the mixed-layer.  This assumption will only be true if the


material is, in fact, well-mixed through the layer and there


is no shear in the layer to disperse the material away from


the mean flow.


     Over sufficiently long travel time (>1.3 hours), the


dispersion of atmospheric admixtures will be dominated by


vertical wind velocity shear.  The climatology of this

-------
                            5.1-30

dispersion for long travel times has been calculated by
Samson (1980) from the divergence of trajectories for sublayers
of the mixed-layer.  The distribution of the probability of
contributing to a measured concentration due to a given mixed
layer trajectory was shown to be normally distributed about
the mixed-layer.  The broadening of the probability field
with length of travel was found to be a function of time.
     The ACID model used this parameterization to describe
the potential contribution of upstream sources ignoring
intermediate chemical transformation, dry deposition and wet
deposition.  The dispersion due to a meander of trajectory
centerlines in times is inherently calculated by sequential
integration of trajectories at six hour intervals over the
period of interest.
     Following the work of Shieh et al., (1979), ACID includes
spatially varying dry deposition with changes in dry deposition
fields from season to season.  To simulate the variation in
dry deposition due to changes in atmospheric stability, the
gridded deposition velocities are forced to vary diurnally.
     The reaction rates associated with the chemical trans-
formation of pollutants will change the probability of being
affected by an upwind source.  Ignoring deposition, a source
close to a receptor may have less chance of producing an
impact on secondary pollutant levels than a source several
hours upstream because of the insufficient time available for

-------
                            5.-1-31

transformation to occur.  ACID assumes a diurnally varying
transformation rate.
     The approach used for wet deposition in the ACID model
is a first order loss rate.  By use of a scavenging coefficient,
the change of probability of contribution is of a simple form.
ACID allows parts of puffs to be washed out without affecting
the remaining region of the puff.  This is accomplished
through a unique technique involving a moving grid following
the course of the trajectory.  An initial probability of
contribution field is calculated at the first time step
upstream (t = 3 hrs) based on dispersion alone.  At each fixed
grid point within the domain of contribution the change in
probability due to dry deposition and chemistry is applied.
Next/ the gridded precipitation field is scanned and the
probability reduced for precipitation areas.
     The location of the fixed grid within the domain is
memorized on a "moving" grid which is moved upstream to the
second time step.  The probabilities are reduced on the moving
grid due to chemical conversion.  .Then the moving grid is
translated to the underlying fixed grid through an efficient
Gaussian smoothing and interpolation scheme whose response
has been designed to simulate the dispersion due to vertical
shear.
     The new fixed grid is again subjected to dry deposition
and wet deposition and the probability is reduced accordingly

-------
                            5. 2-1






and added to the fixed grid.  The location of these fixed



points are then memorized and moved to the third time step.



The total process is then repeated through all time steps.



     The sum of the probabilities can be further summed~over



defined regions and used to define the transfer matrix of




source-to-receptors contributions.



5.2  Intercqmparison and Evaluation



     There is no general agreement in the modeling community



as to the proper method and statistics for intercomparison



and validation of models.   However, some desirable conditions



of the intercomparison process are that input meteorological




and emission data should be common, when possible, and that .



evaluation statistics should be based upon residuals between



simulations and observations, where available.



     Since both meteorological process and emission rates



vary by season, it is instructive to examine simulations for



months and seasons, as well as simulations of annual concen-



trations and deposition.  However, the ability to treat temporal



variations varies among models.



     Model simulations have inherent minimum resolutions, since



there are limitations on input data, as well as internal modeling



limitations.  Observations, on the other hand, are point values.



Since the observations are random realizations of some ensemble



expectation, perfect model/observation correlations can never be



obtained.

-------
                            5.2-2






     The models described in 5.1 have been developed for




different purposes, and no single model is likely to be optimal



for all applications.  The relative value of the models for



particular applications has yet to be determined.



     A preliminary comparison of model estimated average



concentrations and wet depositions has been completed for the



AES-LRT, ASTRAP, MOE-LRT, and RCDM models.



     Comparison of AES, ASTRAP and RCDM monthly  (January and



July) average SC>2 and 804 concentrations and total wet deposit-



ions with monitoring data gave the following results:



     For January 1978, the AES-LRT on average over-predicted



both the S02 and SC>4 concentrations. The ASTRAP model SC>2



concentrations were on average closer to the observed SC>2



concentrations.  These comments are based on only 21 SC>2 and



26 804 data points; hence, statistically there is no discernible



difference between the performance of these two models.  The



RCDM model almost consistently under-predicts SC>2 concentrations



for January 1978 by a factor of 2 or greater; however, January



304 concentration estimates are on the average very close to the



observed concentrations.  The wet sulfur depositions from AES-LRT



and ASTRAP were on the average significantly larger than the



observed depositions.  Predictions of January wet sulfur deposit-



ion by RCDM are on the average close to the observed values.



However,  no overall conclusion can be drawn due to the small



small number of observations (11 stations) and some uncertainty



in the deposition data.

-------
                            5.2-3





     The July 1978 model SC>2 and 804 concentrations are on the



average not as close to the observed as the January estimates



were.  Both AES-LRT and ASTRAP over-predicted the S02 concen-



trations; however, on the average, the ASTRAP simulations for



the 20 sites are close to the observed.  The converse is true



for 804 concentrations (30 stations) in that both models on the



average under-predicted, with AES-LRT simulations closer to the



observed.  RCDM model simulations of July S02 concentrations are



on the average very close to the observed concentrations; how-



ever the RCDM model underestimates the 804 concentrations for



the same period.  July wet sulfur depositions were on the aver-



age slightly over-predicted by the AES-LRT and ASTRAP models and



were on the average under-predicted by the RCDM model.  The



standard deviations for the three models are large and there



were only 15 observations.



     The comparison of the annual model averages with observed



data was possible for only a small data set.  Based on 9 stations



the observed annual SC>2 concentrations on average were over-est-



imated by ASTRAP, AES-LRT and MOE-LRT.  The RCDM model under-est-



imates on the average.  For MOE-LRT and ASTRAP the over-prediction



was less than by the AES-LRT; however, statistically there was no



discernible difference due to the small sample size.  The annual



804 concentrations are on the average under-estimated by the MOE-



LRT and ASTRAP models at the 13 sites considered.  (Note that the



ASTRAP annual average is based on only the arithmetic average of

-------
                            5.2-4





2 monthly averages, namely, January and July).  The AES-LRT model



on the average slightly over-predicts the 804 concentrations.  The



simulations of the RCDM model were on the average very close to



the observed concentrations.



     Annual wet sulfur depositions were slightly over-predicted



by all four models when assessed on an average of the 8 CANSAP



data sites reporting at least 10 months of data in 1978.  The



deviations about the geometric mean"were large for both the ASTRAP



AES-LRT models; however, the MOE-LRT showed much smaller deviat-



ions between the observed and the predicted.  The RCDM predictions



were on the average much lower than the-observed depositions.



     During Phase II a hierachy of model evaluation statistics



were adopted, but only the AES-LRT model computed all of them



(see the AES-LRT Model Profile Report and section 4.2.2 of the



Modeling Subgroup Report).   It is expected that these statistics



will be available from all eight models by September 1981, in an



Addendum to Chapter of the Model Subgroup Interim "Working Re-



port".

-------
                          Chapter 6
             PHASE'II'IMPROVED'EMISSION'INVENTORY
6.1 Emissions
     To eliminate the differences in the emission inventories
used .in the Phase I modeling work, a U.S^-Canada SC>2 emission
inventory was produced bilaterally during Phase II in coopera-
tion with Work Group 3B.  However, the Modeling Subgroup
found it difficult to standardize the source areas in Phase
II because of the re-programming required by some of the
models.  The receptor areas were standardized to an extent
as will be described in the next section.  The inventory
was produced on a state/province, 63 ARMS area and top 50
point source .basis to accommodate the needs of all 5 Phase I
models.
     The best estimates of the current U.S.-Canada SC>2
emissions at the State and Province and the ARMS area levels
are presented in Tables 6-1 and in Table 6-2, respectively.
A list of the top 50 point sources in the eastern U.S. and
Canada is provided in Table 6-3.  The background material to
these data has been incorporated in the Phase II SC"2 Emission
Inventory Report (No. 2-4).  This report supercedes the
information in Appendix 6 and the Addendum to Appendix 6 of
the Phase I report of Work Group 2.
     Work Group 2 is continuing, to work with Work Group 3B to
produce best estimates for primary sulfate, NOX, and historical

-------
                                     6-2

Table 6-1.  Phase II Improved United States and Canadian S02 Emissions
            On a  State and Province Basis (Kilotonnes/Yr)  - 1980
State or Total
Province
Alabama 603.6
Arkansas 100.4
Connecticut 118.3
Delaware 121.8
Dist. of
Columbia 36.1
Florida 839.1
Georgia 580.1
Illinois 1,259.5
Indiana 1,968.0
Iowa 290.0
Kentucky 1,178.3
Louisiana 253.1
Maine 95.9
Maryland 232.0
Mass. 334.6
Michigan 939.6
Minnesota 384.4
Mississippi 205.1
Missouri 1,286.0
New Hamsphire 59.9.
New Jersey 416.4
New York 860.0
North
Carolina 447.9
Ohio 2,907.0
Penn. 1,643.5
Rhode Island 26.3
South
Carolina 260.2
Tennessee 1,034.0
Vermont 13.3
Virginia 386.5
West
Virginia 1,144.9
Wisconsin 604.4
SUB TOTAL 20,612.2
Newfoundland 56.5
Prince Edward
Island 15.8
Nova Scotia 217.3
New Brunswick 223.8
Quebec 1,105.3
Ontario 1,842.8
Manitoba 515.2
SUB TOTAL 3,976.7
Utilities Non-Ferrous
Smelters
443.9
52.9
12.2
87.1

_*
650.5
479.0
892.0
1,475.1
188.7
1,065.0
21.5
21.5
135.1
151.1
674.9
213.7 67.5
147.1
1,096.4 58.4
33.9
187.9
324.3

282.2
2,331.9
986.9 35.3
1.8

173.1
895.8 12.9
0.0
238.2

1,012.0
449.6
14,725.3 174.1"
5.3

5.9
88.2
91.7 12.0
1.6 609.6
436.9 929.8
7.4 485.5
637.0 2,036.9
Transportation
10.3
6.5
7.8
1.3

22.6
16.5
27.6
17.9
9.2
9.4
7.6
2.0
7.5
1.6
13.1
23.0
10.4
5.8
14.1
1.6
19.5
31.4

13.9
35.0
32.0
2.4

7.7
13.9
1.1
14.4

4.9
11.6
4"DTTT"
2.0

0.4
3.1
2.9
18.3
21.3
4.6
52.6
Ind/Res/
Co mm.
95.7
26.8
97.3
20.2 .

10.6
91.2
51.2
250.5
382.4
82.6
80.6
54.3
57.2
76.5
167.6
205.0
75.1
23.7
101.0
22.1
148.1
466.4

109.2
433.1
392.6
21.6

69.2
93.4
11.9
113.2

102.1
131.8
4,'OS87T
41.9

9.5
108.9
100.5
386.2
260.0
13.3
920.3
Other
53.7
14.2
1.0
13.2

2.9
80.9
22.3
99.1
101.3
9.3
25.1
157.3
9.7
24.8
2.8
• 36.7
17.7
28.5
16.1
2.3
60.9
37.9

42.6
107.0
196.7
0.5

10.2
18.0
0.3
20.7

25.9
11.4
1 , 2 51 . d
7.3


17.1
16.7
89.6
194.8
4.4
329.9
* Emissions included with Maryland
  Source of U.S. Data:  Mitre Corporation, April 24, 1981, and E.H. Pechan and
    Associates, May 30, 1981
  Source of Canadian Data:  Environment Canada (Frank Vena)

-------
                                      6-3

Table 6-2.  Phase II United States and Canadian 502 Emissions
            For the 63 ARMS Areas (Kilotonnes/Yr)  - 1980.
ARMS
AREA
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31

UNITED-STATES CANADA
82.6 7.5
. 5.8 12.4
6.7
52.4
355.0
11.2
. 74.1
9.4
220.7
343.3 36.5
611.9
452.1
390.0
675.2
30.1
266.3
125.1
353.6
975.7
148.5
465.0
723.5
675.6
532.3
28.9
274.9
165.9
243.3
429.7
130.2
530.3

ARMS
AREA
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
UNITED-STATES
150.0
27.6
130.2
583.5
200.9
312.3
91.7
3.1
1,289.8
291.5
721.8
402.2
645.5
1,446.0
2,006.1
544.4
339.7
981.8
48.8
611.8
314.8











CANADA

















48.4 .
3.8

0.5
1.2
370.6
872.5
1.3
628.5
18.1
332.7
831.5
5.5
171.9
10.9
Source of U.S. data:  Mitre Corporation, April 24, 1981, and E. H. Pechan
& Associates, May 14, 1981

Source of Canadian data: Environment Canada (Frank Vena)

-------
     Table 6-3.    Combined U.S.-Canadian Top 50 Sources of SC>2 Emissions - 1980
Rank
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Plant
Name
Inco
Noranda
Paradise
Inco
Muskingum River
Gavin
Cumberland
Clifty Creek
Baldwin
Monroe
Labadie
Kyger Creek
Harrison
Johnsonville
Mitchell
Hatfield Ferry
Eastlake
Bowen
Lambton G.S.
Gibson
Nanticoke G.S.
HBMS
Conesville
Shawnee
Algoma Steel
Bruner Land
State/
Province
Ontario
Qjebec
Kentucky
Manitoba
Ohio
Ohio
Indiana
Indiana
Illinois
Michigan
Missouri
Ohio
West Virginia
Tennessee
West Virginia
Pennsylvania
Ohio
Georgia
Ontario
Indiana
Ontario
Manitoba
Ohio
Kentucky
Ontario
Pennsylvania
Emission
(kilotonnes/yr )
807.5
537.5
418.8
333.5
306.7
297.5
296.2
295.3
237.2
224.3
222.6
219.7
215.1
188.0
187.3
173.5
172.8
170.3
160.3
187.8
155.1
152.0
151.8
146.1
143.3
139.3
Rank
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50

i
Plant
Name
Montrose
New Madrid
Sammis
Coffeen
Krammer
Big Bend
Falconbridge Nickel
Keystone
Petersburgh
Conemaugh
Widows Creek
Mount Storm
Cardinal
Stuart
Joppa
Thomas HL
Montour
Kincaid
Gallatin
Gallagher
Gaston
Kingston
Avon Lake
Lake view G.S.


State/
Province.
Missouri
Missouri
Ohio
Illinois
West Virginia
Florida
Ontario
Pennsylvania
Indiana
Pennsylvania
Alabama
West Virginia
Ohio
Ohio
Illinois
Missouri
Pennsylvania
Illinois
Tennessee
Indiana
Alabama
Tennessee
Ohio
Ontario


Emission
(kilotonnes/yr )
138.0
135.1
133.7
124.8
123.8
122.7
122.3
121.5
120.9
118.2
118.8
116.4
115.0
113.1
107.6
101.8
96.9
96.0
95.6
94.6
93.6
92.0
91.7
91.4


Source of U.S. data:  EPA Airtest Program (1979-1980)
Source of Canadian data:  Environment Canada (Frank Vena)

-------
                             6-5.
SOX and NOX emissions on a state/province basis.  In addition,



best estimates for SOX and NOX emissions in the western states



and provinces will be developed in Phase III.



6.2 Source-Receptor Areas



     To generate the Phase II tranfer matrices and evaluate



the models, the subgroup decided to adopt a composite of the



Phase I U.S. and Canadian targeted sensitive areas, (see



Table 6-4).  This provided the two "receptor oriented"



models with the minimum number of Phase II receptor areas.



The other 6 Phase II models ran on either the 11 Canadian



source regions and 9 receptor areas or the full 63 ARMS



areas (all these areas were considered potential receptor



areas as well as source areas).



     In the latter part of Phase II, it was decided that the •



emission source regions to be used in the development of ah



MOI transfer matrix in Phase III be resolved spatially on



a province-subprovince/state-multi-state level, as shown in



Figure 6-1.  Eastern Canada has been subdivided into



15 source areas and the Eastern U.S. into 25 source areas.



These areas are also to serve as receptor areas in Phase III.



The emission and geographic centroids in longitude and latitude



of the Phase III areas are presented in Tables 6-5 and 6-6.

-------
                                   6-6
       Table 6-4.   Phase II Targeted Sensitive Areas
                    For Work Group 2 Modeling
Area                             Longitude and           SURE Grid
Number   Name                  La t i t ude (approx*)     Ce n t r oids (X,Y)

 1       Boundary Waters       93°, 49°                4, 26

 2       Algoma                84°, 46.5°             12, 22

 3       Muskoka               79.5°, 45°             17.5, 19.5

 4       Quebec City           72°, 47°               12.5, 22.5

 5       Southern Nova Scotia  66°, 44°               30.5, 19.5

 6       New Hampshire         72°, 45°               25, 20

 7       Adirondacks           74°, 44°               23.5, 18.5

 8       Western Pennsylvania  78°, 41°               18, '13

 9       Southern Appalachia   84°, 35°               12.5, 5.5

10       Arkansas              92°, 36°               2.5, 7.5

11       Florida               82°, 30°               13.5, -1.5


                          *may vary by +0.5°

-------
6-1. Canadian Province and  Sub-provin<^Hkegions and  U.S.
     State  and Multi-state  Regions for^rie Phase III
     Transfer Matrices
                                                                                           en
                                                                                           i
                                                                          VERMONT
                                                                          NEW HAMPSHIRE
                                                                          MASSACHUSETTS
                                                                          RHODE ISLAND
                                                                          CONNECTICUT
                                                                          NEW JERSEY
                                                                          DELAWARE
                                                                          MARYLAND
                                                                          DISTRICT OF COLUMBIA

-------
                                   6-8
Ible  6-5.   Canadian Regions For Phase III Transfer Matrices

Region

10
11
12
13
14
15
16
77
18
19
20
21
22
23
24
,-,,!,,, 1 . 1 ,,111
Province/
Sub-Province

Northern Manitoba
Southern Manitoba
Northern Ontario
West
Northern Ontario
East and
Algoma South
Sudbury
Southwestern
Ontario
Southeastern
Ontario
St. Lawrence
Valley - Quebec
Noranda and
' North-Central
Quebec
Gaspe Bay-Quebec
New Brunswick
Nova Scotia and
Prince Edward
Island
Newfoundland and
Labrador
Saskatchewan and
Alberta
British Comumbia
'
Area
Lat.

56.84
51.21
52.02
49.03
46.32
43.56
45.23
46.88
54.28
50.02
46.21
45.46
51.43
50.00*
49.30*

Centroids
Long.

97.77
98.34
89.20
83.03
80.70
81.24
77.40
72.13
73.59
64.56
66.77
63.27
58.89
110.00*
123.00*

Emission
Lat.

55.61
49.97
49.32
48.05
-
43.39
45.07
46.24
50.20
49.35
45.75
45.52
49.12
50.00*
49.30*

Centroids
Long.

98.25
98.72
90.00
83.76
-
80.59
75.38
72.70
74.02
66.14
65.98
62. 57
56.08
110.00*
123.00*
lay be revised

-------
                                 6-9
ble 6-6.  U.S. Regions For Phase III Transfer Matrices
Emissions Centroids Area Centroids
R,eg.ipn
50
51
52
53
54
55
56
57
58
59
r
61
62
63
64
65
66
67
68
69
70
71
72
1,
s.ta.te
Ohio
Illinois
Pennsylvania
Indiana
Kentucky
Michigan
Tennessee
Missouri
West Virginia
New York
Alabama
Wisconsin
Iowa
Minnesota
Virginia
North Carolina
Florida
Georgia
South Carolina
Maryland
Delaware
New Jersey
Dist. of Col.
Arkansas
,LpAg,«- , La,t,., ,;.,.,
82.5
88.0
79.0
87.0
85.5
84.0
86.0
91.5
80.5
74.0
87.0
88.5
92.0
93.5
77.8
79.0
82.0
84.0
80.8
76.3
75.3
74.4
77.0
92.5
40.0
41.5
40.8
40.0
37.5
43.0
35.9
38.8
38.5
42.0
34.0
43.5
42.0
45.0
37.8
36.0
29.0
33.5
33.8
39.2
39.1
40.5
38.9
34.8
Lpng...
82.5
89.0
77.7
86.1
85.5
84.5
86.0
92.5
80.5
76.0
86.8
89.7
93.5
94.5
78.8
79.0
82.0
83.5
80.8
76.3
75.3
74.5
77.0
92.5
,Lat,.
40.2
40.0
41.0
40.0
37.5
44.5
35.9
38.5
38.5
42.7
33.0
44.5
42.0
46.2
37.8
35.5
29.0
33.0
33.8
39.0
39.1
40.2
38.9
34.8

-------
                                6-10

 Table 6-6 (continued).

                                   Emissions Centroids    Area Centroids
BBK • 'j Vi
74
75
76
77
78
79
80
81
82
83
84
i
^^^
86
87
88
89
90
91
92
93
94
95
96


i S,t.a.t.e
Louisiana
Mississippi
Massachusetts
Connecticut
Rhode Island
Maine
Vermont
New Hampshire
North Dakota
South Dakota
Nebraska
Kansas
Oklahoma
Texas
Montana
Wyoming
Colorado
New Mexico
Idaho
Utah
Arizona
Washington
Oregon
Nevada
California
jlofig.' Lat. r' '' JLojrcg. Lat.' '
91.0 . 30.5 92.0 31.0
89.5 33.0 89.5 32.5
71.5 41.7 72.0 41.7
72.7 41.5 72.7 41.7
71.5 41.7 71.6 41.7
69.0 44.5 69.0 45.5
73.0 44.0 72.7 43.9
71.5 44.0 71.5 44.0
99.0 47.0 100.3 47.0
100.0 44.5 100.0 44.5
97.0 41.0 100.0 41.5
97.0 38.5 98.3 38.5
96.8 36.0 97.8 35.5
97.0 31.0 99.0 32.0











Note:  The appropriate centroids for states 39-49 will be determined
       in Phase III

-------
                            Chapter 7

                7.  SOURCE-RECEPTOR RELATIONSHIPS
                         "FOR"SULFUR OXIDES
7.1  Introduction

     Several long-range transport models are currently available

for estimating annual sulfur deposition and for studying source-

receptor relationships.  These models ae briefly described in

Chapter 5 and more completely in Chapter 2 of the Modeling Subgroup

Report.  Some preliminary attempts to model nitrogen oxides are

discussed in Chapter 9.

     Eastern North America has been subdivided in several ways for

studying source-receptor relationships as described in Chapter 4 of

the Atmospheric Modeling (Work Group 2) Phase I Report.  In Phase

II, all models retained their original format for emissions data

input and source-receptor transfer matrix elements were determined

for the Canadian 11 source regions and 9 Phase I targeted sensitive

areas on an annual basis (see Figure 7-1 and Table 7-1).  The models

used at least January and July 1978 meteorological data and the

Phase II sulfur dioxide emissions inventory in Chapter 6 of this

report.

7.2  Source-Receptor Matrix Description

     In both Phases I and Phase II, annual S02 and 804 concen-

tration, and dry, wet and total sulfur desposition transfer matrices

were generated by the MOE, AES, ASTRAP, ENAMAP, and RCDM models.

-------
7-2
           Figure 7-1.   Map of East-
           ern North America Showing
           the 11 Major Source Regions
           (10 shaded and the rest
           east of the  Mississippi
           River) and 9 Sensitive
           Areas Used in the Phase II
           Transfer Matrices

-------
                               7-a
Table 7-1.  Key to 11 Source Regions and 9 Sensitive Areas
     Source Regions



1    Michigan                          1



2    Illinois - Indiana                2



3    Ohio                              3



4    Pennsylvania                      4



5    New York to Maine                 5



6    Kentucky-Tennessee                6



7    West Virginia to North Carolina   7



8    Rest of Eastern U.S. (Florida     8



9    Ontario                           9



10   Quebec



11   Atlantic Provinces
Sensitive Areas



Boundary Waters



Algoma



Muskoka



Quebec



Southern Nova Scotia



Vermont-New Hampshire



Adirondacks



Western Pennsylvania



Smokies

-------
                               7-4
Since in Phase I, the ENAMAP Model domain did not inrlndp th^



Atlantic provinces of Canada, the effects of emissions from this



region were not considered by the model.



     In Phase I, the modelers did not use the same meteorological



periods and emissions inventory.  In order to provide a means



whereby the matrix element values computed by each of the



models could be intercompared, the values were normalized using a



unit emissions rate of 1 Tg of sulfur per year.  The normalized



values for annual wet sulfur deposition (kg.S.ha.~lyr.~1) generated



during Phase I are presented in Table 7-2.  The actual sulfur



emission rates used for each region by each model are presented



in the third column of this table.



     The matrix element with the greatest variation among models



is that for the impact of the Ontario source region (9) on Muskoka



(3).  The values ranged from 1.60 (MOE) to 12.86 (ENAMAP) kg.S.



ha.'^yr."1 per Teragram of sulfur emission.  However, the



large value calculated by the ENAMAP model was probably due to



the location of the Muskoka receptor area about 250 km closer to



the Sudbury, Ontario, point source than in the other models.



     Other matrix elements where the values varied significantly



were the impact of the Pennsylvania source region (4) on the



Pennsylvania sensitive area (8) [4.24(RCDM) to 10.61 (ENAMAP)];

-------
Table 7-2.  Phase I Transfer Matrix of:

            Annual Wet Deposition of Sulfur (kg.ha~l.yr~l)
            per unit emission (Tg.S.yr~l)
(1)

Source
Regions
1
Mich.
2
111.
Ind.
3
Ohio
4
Penn.
5
N. York
to Maine
6
Kent.
Term.

Models
MOE
AES
ASTRAP
ENAMAP
RCDM
MOE
AES
ASTRAP
ENAMAP
RCDM
MOE
AES
ASTRAP
ENAMAP
RCDM
MOE
AES
ASTRAP
ENAMAP
RCDM
MOE
AES
ASTRAP
ENAMAP
RCDM
MOE
AES
ASTRAP
ENAMAP
RCDM

Erniss.
(Tg.S)
0.784
0.973
1.194
1.194
1.194
2.538
1.937
2.077
2.077
2.077
1.983
2.381
2.163
2.163
2.163
1.021
1.028
0.990
0.990
0.990
1.143
1.204
1.208
1.208
1.208
1.202
1.418
1.473
1.473
1.473









B. Waters
(1)
0.07
0.21
0.44
0
1.11
0.06
0.05
0.24
0
0.79
0.04
0
0.08
0
0.30
0.03
0
0.01
0
0.10
0.02
0
0
0
0.05
0.03
0
0.07
0
0.28

Alg.
(2)
0.40
2.40
2.59
0.97
3.95
0.23
1.20
1.32
0.11
1.14
0.15
0.25
0.86
0.06
0.80
0.12
0.29
0.29
0
0.36
0.07
0.17
0.08
0
0.21
O.lO
0.14
0.45
0.01
0.38

Musk.
(3)
0.93
3.20
2.50
1.67
2.51
0.32
1.10
1.40
0.08
1.35
0.32
1.80
1.84
0.36
2.32
0.28
1.30
1.13
0.04
1.62
0.19
0.50
0.44
0.01
1.08
0.14
0.71
0.76
0.01
0.63
, Recei
Que.
(4)
0.34
1.00
1.46
0.13
1.68
0.15
0.31
0.75
0.02
0.81
0.19
0.46
1.04
0.02
1.23
0.21
0.68
1.10
0
1.06
0.25
1.30
0.79
0
0.99
0.09
0.07
0.42
0
0'.36
5tor Areas
S. N.Sc.
. (5) .
0.39
0.31
0.55
0.19
0.38
0.18
0.10
0.37
0.02
0.26
0.28
0.21
0.60
0.09
0.52
0.40
0.29
1.20
0.07
0.95
1.00
2.00
2.51
2.86
2.33
0.13
0.07
0.24
0.01
0.17

Vt. NH.
(6) .
0.56
0.72
0.73
0.20
0.75. .,
0.23
0.30
0.52
0
0.48
0.32
1.00
0.91
0.05
0.96 .
0.39
1.80
1.79.
0.07
1.59
0.56
2.20
3.00
1.28
2.98
0.14
0.21
0.38
0.01
0.29

Adir.
(7).
0.86
1.10
1.14
0.11
1.11
0.31
0.36
0.82
0.01
0.74
0.47
1.30
1.50
0.12
1.60
0.57
2.20
2.59
0.32
2.63
0.80
2.40
2.50
0.97
3.19
0.18
0.42
0.63
0.01
0.46

Penn.
(8) .
1.70
1.70
0.81
0.13
1.09
0.76
1.10
1.18
0.06
1.40
2.00
4.70
3.29
1.01
3.66
4.40
7.90
4.61
10.61
4.24
0.33
0.42
0.69
0.07
0.88
0.46
1.50
1.88
0.13
1.48

Smokies
(9)
0.12
0.21
0.07
0.14
0.32
0.47
0.77
0.60
0.36
0.91
0.23
0.25
0.53
0.09
0.83
0.11
0.10
0.07
0
0.43
0.05
0
0.01
0
. 0.13
1.60
3.10
4.22
4.24
3.95
                                                      01

-------
Table 7-2. Phase I Transfer Matrix of: (continued)
           Annual Wet Deposition of Sulfur (kg.ha
           per unit emission (Tg.S.yr~l)
7
W.Virg.
to N.C.



8
Rest of
(USA) Fid
to Mo. to
Minn.

9
Ontario



10
Quebec



'"11 	
Atlantic
Provinces




MOE
AES
ASTRAP
ENAMAP
RCDM

MOE
AES
ASTRAP
ENAMAP
RCDM
MOE
AES
ASTRAP
ENAMAP
RCDM
MOE
AES
ASTRAP
ENAMAP
RCDM

MOE
AES
ASTRAP
ENAMAP
RCDM

1.703
1.223
1.610
1.610
1.610

1.196
3.743
4.012
4.012
4.012
0.906
0.985
0.949
0.949
0.949
0.595
0.519
0.464
0.464
0.464

0.187
0.235
0.453
0.453
0.453





























0.03
0
0.02
0
0.14

0.09
0.24
1.09
1.90
2.16
0.08
0.10
0.09
0
0.29
0.06
0
0.02
0
0.02

0.01
0
0
	 *
0

0.08
0
0.34
0.01
0.35

0.39
0.61
0.56
0.46
0.92
0.51
1.80
2.92
1.32
2.04
0.18
0.19
0.52
0
0.14

0.03
0
0.02
—

0.15
0.33
0.80
0.11
1.08

0.34
0.24
0.44
0.15
0.42
1.60
3.30
4.71
12.86
5.07
0.32
0.58
1.14
0
0.47

0.05
0
0.10
—
0.02 | 0.06

0.13
0.33
0.61
0
0.65

0.15
0.05
0.29
0.03
0.32
1.00
1.70
3.83
0.59
4.10
1.50
2.90
2.61
0
1.23

0.16
0.43
0.63
—
0.11

0.28
0.25
0.56
0.02
0.52

0.15
0.03
0.15
0.03
0.08 . .
0.57
0.61
1.17
0.20
0.86
0.73
0.96
2.06
2.20
1.50 .

0.74
2.60
3.07
—
2.42

0.22
0.90
0.87
0.05
0.79

0.20
0.08
0.20
0.02
0.15 . .
1.10
1.60
1.41
0.16
1.74
2.30
3.30
2.09
1.50
2.41 .

0.16
0
2.48
—
0.60

0.29
1.10
1.24
0.06
1.23

0.26
0.13
0.27
0.02
0.23
1.20
2.00
3.08
0.63
2.51 ,
0.59
1.50
1.90
0.04
1.33

0.10
0
1.23
—
0.23

0.85
3.50
4.04
2.67
4.99. .

0.40
0.53
0.50
0.31
0.46
0.53
1.20
0.43
0.18
0.90. .
0.13
0.19
0.12
0
. 0.15

0.05
0
0.08
—
0.04

0.16
0.49
1.18
0.26
1.39

1.00
2.50
1.40
0.14
1.21 .
0.05
0
0.02
0.01
0.14
0.03
0
0
0
0.02

0.01
0
0
—
0.01
                                                                                                                 cr\
   ENAMAP did not consider source region 11 in the Phase I modeling
(1)  Annual deposition transfer matrix elements for the ASTRAP and ENAMAP models are based on an
     arithmetric mean of the equivalent January 1978 and July 1978 elements.

-------
                               7-7




Michigan (1) on Algoma (2) [0.40 (OME) to 3.95 (RCDM)]; Ontario



(9) on Quebec (4)  [0.59 (ENAMAP) to 4.10 (RCDM)]; Ohio (3) on



Pennsylvania (8) [1.01 (ENAMAP) to 4.70 (AES)]; and Virginias and



North Carolina  (7)  on Pennsylvania (8) [0.85 (MOE) to 4.99(RCDM)].



     In Phase II, several changes in the input data and the para-



meterizations in some of the models resulted in revisions of the



values of the transfer matrix elements.  These changes included



the use of:



     1)  the same meteorological periods (wind and



         precipitation data for January and July 1978),



     2)  the same sulfur emissions inventory (1 Tg sulfur



         from each  source region)



     3)  the same locations of receptor areas (after



         Phase  I, it was discovered that not all the



         modelers were using the same location for several



         receptor areas)



     4)  a more complex and realistic wet and dry sulfur



         deposition parameterization scheme in the



         ENAMAP model.



     5)  a change in the relative percentage of wet and dry



         sulfur deposition in the RCDM model



With the use of standardized input data,  one would expect the



variations in each  set of transfer matrix values to decrease



somewhat.

-------
                               7-8



     Table 7-3 presents the normalized (1 Tg sulfur per year)



Phase II annual wet sulfur deposition transfer matrix values.



The complete set of Phase II transfer matrices are presented



in Appendix C of the Phase II Modeling Subgroup report.



7.3  Intercomparison of Phase I and Phase II Transfer Matrices



     In this section, selected  Phase I and Phase II annual



wet sulfur deposition transfer matrix element•values are inter-



compared.  In addition, the 1978 wet sulfur deposition calculated



by the models at each targeted receptor area are compared to



measured values.



     Table 7-4 presents selected transfer matrix element values



computed in Phase I and Phase II.  The values in these elements



varied the greatest in Phase I.



     For three of these six source-receptor pairs (Ontario-Muskoka,



Pennsylvania-Pennsylvania, and Virginias and North Carolina-



Pennsylvania) the maximum value and the range of values of wet



sulfur depositon increased.  The range of values in the other



three source-receptor pairs decreased.  The maximum value decreased



markedly in Phase II for the Michigan-Algoma (from 3.95 to 2.40)



and Ontario-Quebec  (from 4.10 to 2.16) source-receptor pairs.



The deposition parameterization in both the RCDM and ENAMAP



models were modified significantly in Phase II.  The ASTRAP model



was not modified, but it was rerun using the standardized meteoro-



logical and emissions data sets.

-------
Table
Phase II Transfer Matrix of:
Annual Wet Deposition of Sulfur (kg.ha~l,
per unit emission (Tg.S.yr~l)


Source
Regions
1
Mich.





2
111.
Ind.




3
Ohio





4
Perm.





5
N. York
to Maine






Models
MOE
AES
ASTRAP
ENAMAP
RCDM
MEP
MCARLO
MOE
AES
ASTRAP
ENAMAP
RCDM
MEP
MCARLO
MOE
AES
ASTRAP
ENAMAP
RCDM
MEP
MCARLO
MOE
AES
ASTRAP
ENAMAP
RCDM
MEP
MCARLO
MOE
AES
ASTRAP
ENAMAP
RCDM
MEP
MCARLO

Emiss.
(Tg.S)
0.784
0.973
1.194
1.194
1.194


2.538
1.937
2.077
2.077
2.077


1.983
2.381
2.163
2.163
2.163


1.021
1.028
0.990
0.990
0.990


1.143
1.204
1.208
1.208
1.208


Receptor Areas
B. Waters

(1)
0.07
0.21
0.06
0.04
0.30
0.14
0.06
0.06
0.05
0.07
0.00
0.25
0.05
0.01
0.04
0.00
0.02
0.00
0.11
0.01
0.00
0.03
0.00
0.00
0.00
0.04
0.01
0.00
0.02
0.00
0.00
0.00
0.02
0.00
0.00
Alg.

(2)
0.43
2.36
1.47
0.71
0.94
1.24
0.50
0.23
1.24
0.61
0.11
0.37
0.32
0.23
0.16
0.25
0.31
0.00
0.31
0.14
0.08
0.13
0.29
0.14
0.00
0.16
0.08
0.03
0.08
0.17
0.02
0.01
0.10
0.05
0.01
(1) Annual deposition matrix elements for AST
Musk.

(3)
0.92
3.19
3.44
1.40
1.37
1.27
0.61
0.31
1.14
1.02
0.58
0.50
0.31
0.28
0.32
1.85
0.75
0.02
0.91
0.95
0.33
0.29
1.26
0.48
0.00
0.84
0.60
0.20
0.22
0.50
0.12
0.04
0.53
0.66
0.17
Que.

(4)
0.33
1.03
1.10
1.32
0.73
0.18
0.33
0.15
0.31
0.71
0.59
0.28
0.08
0.20
0.18
0.46
0.92
0.75
0.47
0.23
0.33
0.21
0.68
0.74
0.73
0.54
0.45 '
0.34
0.26
1.33
1.38
0.60
0.47
1.24
0.35
S. N.Sc.

(5)
0.37
0.31
0.20
0.00
0.28
0.08
0.06
0.18
0.10
0.12
0.00
0.12
0.02
0.04
0.28
0.21
0.25
0.00
0.23
0.08
0.05
0.39
0.29
0.52
0.00
0.37
0.20
0.05
0.98
1.99
0.59
0.00
0.53
1.06
0.08
Vt. NH.

(6)
0.54
0.72
1.03
0.73
0.35
0.25
0.31
0.22
0.26
0.70
0.46
0.17 .
0.12
0.17
0.31
1.01
0.83
0.99
0.38
0.33
0.27
0.38
1.75
1.94
1.55
0.79
0.83
0.36
0.59
2.24
3.02
1.63
1.30
2.29
0.52
Adir.

(7)
0.83
1.13
2.33
1.35
0.57
0.42
0.40
0.30
0.36
1.16
0.35
0.28
0.13
0.28
0.46
1.34
1.79
0.90
0.65
0.47
0.47
0.57
2.24
2.15
3.91
1.37
1.28
0.50
0.83
2.41
2.94
3.46
1.71
1.89
0.29
Penn.

(8)
1.66
1.75
1.08
1.56
0.69
0.66
0.2*7
0.74
1.14
1.04
1.12
0.66
0.17
0.32
1.99
4.75
3.91
7.24
1.88
1.97
0.61
5.42
7.88
12.16
12.82
1.37
6.46
0.68
0.40
0.42
0.31
3.37
0.46
0.63
0.22
Smokies

(9)
0.12
0.10
0.07
0.79
0.18
0.11
0.20
0.47
0.77
0.40
0.56
0.54
0.38
0.43
0.24
0.25
0.24
0.23
0.42
0.29
0.23
0.12
0.00
0.04
0.24
0.13
0.16
0.05
0.05
0.00
0.01
0.06
0.06
0.01
0.01
[•RAP and ENAMAP models are based on the mean of the
equivalent January 1978 and July 1978 elements.

-------
Table 7-3. Phase II Transfer Matrix of:  (continued)
           Annual Wet Deposition of Sulfur  (kg.ha~-"-.yr~l]
           per unit emission (Tg.S.yr~l)
6
Kent.
Tenn.




7
W.Virg.
to N.C.




8
Rest of
(USA) Fid
to Mo. to
Minn.


9
Ontario





10
Quebec





11
Atlantic
Provinces




MOE
AES
ASTRAP
ENAMAP
RCDM
MEP
MCARLO
MOE
AES
ASTRAP
ENAMAP
RCDM
MEP
MCARLO
MOE
AES
ASTRAP
ENAMAP
RCDM
MEP
MCARLO
MOE
AES
ASTRAP
ENAMAP
RCDM
MEP
MCARLO
MOE
AES
ASTRAP
ENAMAP
RCDM
MEP
MCARLO
MOE
AES
ASTRAP
ENAMAP
RCDM
MEP
MCARLO
1.202
1.418
1.473
1.473
1.473


1.703
1.223
1.610
1.610
1.610


1.196
3.743
4.012
4.012
4.012


0.906
0.985
0.949
0.949
0.949


0.595
0.519
0.464
0.464
0.464


0.187
0.235
0.453
0.453
0.453












































0.03
0.00
0.03
0.00
0.14
0.01
0.00
0.03
0.00
0.00
0.00
0.06
0.01
0.05
0.09
0.24
0.24
0.12
0.37
0.23
0.06
0.08
0.10
0.04
0.02
0.17
0.07
0.20
0.06
0.00
0.00
0.00
0.03
0.03
0.01
0.01
0.00
0.00
0.00
0.00
0.00
0.00
0.10
0.14
0.22
0.00
0.18
0.05
0.04
0.08
0.00
0.05
0.00
0.15
0.02
0.01
0.38
0.61
0.38
0.15
0.25
0.34
0.11
0.62
1.83
5.43
1.45
1.30
2.17
0.47
0.20
0.19
0.19
0.00
0.12
0.34
0.09
0.03
0.00
0.00
0.00
0.01
0.01
0.00
0.14
0.71
0.40
0.03
0.28
0.09
0.10
0.16
0.33
0.25
0.00
0.51
0.16
0.09
0.33
0.27
0.37
0.02
0.18
0.12
0.09
1.63
3.35
6.89
1.41
1.91
2.81
0.61
0.38
0.58
0.65
0.00
0.30
0.49
0.28
0.06
0.00
0.00
0.00
0.04
0.06
0.01
0.09
0.07
0.14
0.05
0.16
0.03
0.11
0.13
0.33
0.37
0.38
0.30
0.09
0.14
0.15
0.05
0.12
0.08
0.12
0.02 '
0.05
1.00
1.73
2.15
1.61
1.52
0.72
0.44
1.58
2.89
7.71
.0.90
0.59
3.53
0.56
0.18
0.43
0.02
0.07
0.10
0.30
0.05
0.13
0.07
0.04
0.00
0.07
0.01
0.01
0.27
0.25
0.26
0.00
0.18
0.05
0.02
0.15
0.03
0.03
0.00
0.05
0.01
0.01
0.55
0.61
0.16
0.00
0.56
0.14
0.08
0.71
0.96
0.25
0.00
2.33
0.57
0.21
0.75
2.55
0.87
0.00
0.93
1.42
0.17
0.13
0.21
0.52
0.12
0.11 .
0.04
0.07
0.22
0.90
0.84
0.83
0.37
0.22
0.15
0.20
0.08
0.15
0.12
0.07
0.03
0.05
1.04
1.62
1.18
1.20
0.64
0.89
0.37
2.99
3.28
'1.54
3.06
0.81
3.15
0.20
0.19
0.00
0.03
5.78
0.33
0.19
0.04
0.17
0.42
0.96
0.17
0.19
0.06
0.16
0.29
1.14
1.03
1.77
0.60
0.28
0.20
0.25
0.13
0.26
0.15
0.11
0.05
0.09
1.14
2.03
2.18
1.75
0.95
1.56
0.32
0.65
1.54
2.11
1.35
0.53
1.87
0.17
0.12
0.00
0.00
5.39
0.12
0.10
0.01
0.45
1.48
1.30
1.16
0.61
0.18
0.26
0.91
3.52
5.98
7.73
2.79
1.25
0.42
0.39
0.53
0.32
0.19
0.25
0.05
0.08
0.54
1.22
0.25
0.53
0.39
0.99
0.17
0.13
0.19
0.12
0.03
0.13
0.16
0.10
0.05
0.00
0.00
0.35
0.02
0.07
0.00
1.64
3.10
8.17
3.91
1.29
2.01
0.75
0.18
0.49
0.13
2.73
0.48
0.75
0.17
0.99
2.46
2.74
3.05
0.81
0.29
0.20
0.05
0 00
0.02
0.25
0.06
0.05
0.09
0.03
0.00
0.00
0.03
0.03
0.00
0.01
0.02
0.00
0.00
0.09
0.01
0.00
0.00
                                                                                                                   H-1
                                                                                                                   o

-------
                                   7-11
      Table 7-4.  Selected Transfer Matrix Element Values of Annual
                  Wet Sulfur Deposition from Phase I and Phase II
Source-Receptor
        Range of
     Phase I  Values
(kg.S.ha.~!yr.~1/Tg.S.yr.
      Range of
    Phase II Values
(kg.S.ha.'lyr.'VTg.S.yr."1)
Ontario-Muskoka

Pennsylvania-
 Pennsylvania

Michigan-Algoma

Ohio-Pennsylvania

Virginias and
 North- Carolina-
 Pennsylvania

Ontario-Quebec
     1.60 -  5.07


     4.24 - 10.61

     0.40 -  3.95

     1.01 -  4.70



     0.85 -  4.99

     0.59 -  4.10
      1.60 -  6.90


      1.37 - 12.17

      0.40 -  2.40

      1.88 -  4.70



      0.85 -  5.98

      0.72 -  2.16

-------
                               7-12






     The wet sulfur deposition transfer matrix element values



generated by the RCDM model decreased by an order of magnitude.



The largest RCDM and ASTRAP Phase I values and the corresponding



Phase II values are "listed in Table 7-5.



     Each of the RCDM Phase II values decreased markedly to 1/2



l/3rd their Phase I value).  On the other hand, the ASTRAP Phase



II values tended to increase from their Phase I values, due



primarily to the difference in meteorological periods.  (i.e. the



increase from 4.61 to 12.17 in the Pennsylvania - Pennsylvania



relationship.



7.4  Comparison of Model Estimates with Observations



     Individual transfer matrix elements cannot be directly verified,



but total wet deposition can be compared to observations.  Some



observations of wet sulfur deposition have been reported and have



been compared with model estimates at these sites (see Phase I Report)



Table 7-6 is an enlargement of that data comparison showing the same



five Phase I and Phase II models.



     Although the Phase I model estimates show a wide variation at



some of the sensitive areas, they all show a maximum value at the



Pennsylvania site and have an average value of about 25.0 kg.S.ha."!



yr.'1 and a standard deviation of 6.4 kg.S.ha.-1yr."!.  The average



value overestimates by about 30% because of the tendency of some



models to overestimate by 30% because of the tendency of some

-------
                               7-13
Table 7-5.
Significant Changes in Phase I and II
RCDM and ASTRAP Annual Wet Sulfur Depositions


Model

RCDM







ASTRAP









Source- receptor

Ontario-Muskoka
Virginias and NC-Penn
Ontario-Quebec
Michigan-Algoraa
Ky/Tenn- Smokies
Ohio-Penn
N. York to Maine-
Adir.
Ontario-Muskoka
Penn-Penn
Ky/Tenn-Smokies
Virginias and NC-Penn
Ontario-Quebec
Ohio-Penn
Ontario-Adir
Ontario-Algoma
Phase I Wet
Sulfur Deposition
(kg ha'1 yr -1 )

5.07
4.24
4.10
3.95
3.95
3.66

3.19
4.71
4.61
4.22
4.04
3.83
3.29
3.08
2.92
Phase II Wet
Sulfur Deposition
t (kg ha -1 yr'1)

2.79
•1.91
1.52
0.94
1.29
1.88

1.71
6.90
12.17
8.17
5.98
2.16
3.92
2.19
5.43

-------
         Table 7-6.   Preliminary Model Estimates and Observations of Annual Wet Sulfur Deposition
                     (kg.S.ha.~^yr~l)  at  the Nine Targeted Sensitive Areas


                                      Phase  I                                 Phase II
Sensitive Obs.l1'
Areas Values
1 B. Waters 3
2
3
4
5
6
7
8
9
Algoma
Muskoka
Quebec
S.N. Scotia "
Vt., N.H.
Adirondacks
Pennsylvania
Smokies
3
6
5
3
7
7
12
7
1 Canadian
OME*
3
5
7
6
7
8
8
17
7
AES
2
10
18
9
6
13
16
34
17
United States
RCDM
16
20
23
17
10
15
19
27
18
ASTRAP
7
16
21
17
12
15
19
25
17
ENAMAP
10
6
19
1
11
4
3
21
8
Canadian
OME*
3
5
7
6
7
8
8
17
7
AES
2 .
10
18
9
6
13
16
34
17
MEP
1
6
10
6
3
8
9
16
7
United
States
ASTRAP ENAMAP
2 0.7
17
25
24
5
21
31
52
36
5
7
13
0
19
27
71
35-



RCDM MCARLO
4 0.8
8
15
10
7
9
1-3
21
12
3
5
5
1
5
6
7
6




(-•
**


*Background of 2 Kg.S.ha.~1yr.~  added to MOE values


(1)  Source:  Interpolated from a map prepared by the National Atmospheric Deposition
              Program (1981)  based on data for March 1979 to March 1980.

-------
                                7-15



models to overestimate deposition near large emission regions.


The models show an average annual wet sulfur deposition rate of 13


kg.S.ha.~%r."^ at the Adirondacks sensitive area, which


compares very well to the observations.  The standard deviation of


the model results is 7.2 kg.S.ha.~%r."^ which reflects the


wide range in model values.  Table 7-6 shows the Phase I and II


estimates by the OME-LRT and AES-LRT models are the same while the


estimates by the ASTRAP, ENAMAP and RCDM models changed significant-


ly due to the use of 1978 meteorology and the Phase II S02 emissions


inventory and some changes to the input parameters.  In general/


the Phase II-estimates by the OME-LRT, MEP-TRANS and MCARLO models

           +
and within -50% of the observed values, while the estimates by the
                                   +
AES-LRT and RCDM models are within -75% of the observed values.


The ASTRAP and ENAMAP model estimates are generally much higher


than the observed values at all the targeted sensitive areas.  All


the models show a general tendency to over-predict the observed


wet sulfur depositions in the targeted sensitive areas except for


the MCARLO model.  These preliminary evaluation results will be


used by the modelers to refine their input parameters and check


their results before starting the second round of model evaluation


(see Section 6).

-------
                          Chapter 8






       ANALYSES OF TRANSFER MATRICES FOR SULFUR OXIDES
8.1  Introduction



     Annual transfer matrices from the different LRTAP models



were displayed in Chapter 7.  These matrices represent mean



source-receptor relationships calculated by the models for



various meteprogical periods and then normalized and averaged



to be expressed as annual matrices.  When computing the



matrices/ the values of the model parameters were chosen



independently for each model.  Also/ no attempt was made to



"tune" the models with measurements to obtain the "best fit"



parameters.



     Before using transfer matrices in the assessment iterations/



it is necessary to determine the seasonal variability and



sensitivity of the transfer matrices to uncertainties in the



model parameters.  In this chapter, some preliminary results



on seasonal variability obtained with the ASTRAP and the



AES-LRT models will be presented in the next section.  In



section 8.3, sensitivity to variations to input parameters



are examined for the AES-LRT, OME-LRT and the MEP-TRANS models.



8.2  Seasonal Variations
8.2.1  AES-LRT Results



     The annual transfer matrices from the AES-LRT model were



computed using meteorological data for all of 1978 and the

-------
                             8-2





parameter values in Table 8-1.  The parameter values,



taken as annual averages, were held constant throughout the



year, while in reality the parameters vary spatially and



temporally.



     To test the adequacy of simulating the 1978 annual



concentrations and depositions with average parameter values,



the quantities were recomputed using seasonally varying



parameters.  The deposition velocities and transformation



rates were increased during summer season.  During the winter,



the washout ratio for sulfur dioxide was increased, while the



values for deposition velocity of sulfur dioxide, washout



ratio for sulfate and transformation rate were decreased as



shown in Table 8-1.  The average parameter values were, assumed



to apply during fall and spring.  Emission rates were held



constant throughout the year.



     The differences between concentrations and depositions/



computed with the temporally varying parameters and fixed



parameters were site dependent, but small (see Tables 8-2



and 8-3).  Concentrations of sulfur dioxide and depositions



of wet and dry sulfur for the two computations were within



10 per cent of one another.

-------
                                          8-3
Table 8-1. AES-LRT Model Parameter Values Used
           in the Sensitivity Study
Parameter

Dry deposition
VDSO (cm s"1;
Dry depositior

velocity of
1
i velocity of

S02
S04
Base Case
0.5
0.1
. Winter
0.2
0.1
Summer
1.5
0.6

 VDS04) (cm s


 Washout ratio S02:WS02 (dimensionless)     3.0'IQ4          3.0*105      3.0*104


 Washout ratio S04:WS04 (dimensionless)     8.5'105          8.5'104       8.5*105


 Transformation rate: k (% hr"1)            1.0              0.3           1.5

 Mixing Height: H(m)                        (1)              (2)           (3)
 (1):  climatological monthly mean values for the entire year

 (2):  climatological monthly mean values for the season indicated

-------
                                          8-4

Wable 8-2.   Seasonal Variations  in  the AES-LRT Model Transfer Matrices
             of Absolute Values  (1978)
                         WET DEPOSITION OF SULFUR (kqS.ha."1)

PERIOD
Spring
Summer
Fall
Winter
Composite
Year
^se Year .
Per Cent
Difference

. BNDW.
0.43
. , .0.30 .
0.22
0.31
1.26
1.50
16

ALG.
3.66
. 1.46 .
2.88.
1.76
9.76
10.36
6

. MUSK.
. . 4.46 .
, 2.87
5.46
3.69
16.48
17.6.0
6
. . I
QUE.
2.01
1.31
2.37
2.29
7.98
9.02
12
RECEPTOR
. S.NSC.
0.98
.0...53
, 1.89.
.2.20
. 5.60
5.59 .
. 0

VT-NH .
2.57. ,
. 1.87
. 4.18. .
4.24.
12.86
13.10
2

ADIR. .
. 2.97 .
.2.18. .
, 5.02,
., 4.90
, 15.07
15.74 .
.4 . .

PENN.
6.13
. 7.42
9.24 .
13.06
35.86
33.51
-7

SMOKIES
. 3.92
2.39 .. ,
. 3.4.9 . .
. 8.11
1.7.91
16.66
. -8 . .
                             Average S09 CONCENTRATION   Qag/m3)

PERIOD
Spring
Summer
Fall. .
Winter ...
Composite
Year
fcase Year
Per Cent
Difference

BNDW..
.4.2
1.2
., 2.3 . ,
. . 3.8,
2.9
2.9
2

ALG.
12.7 .
3.6
.1.0.1 .
1.1.3 .
9.4
10.3 .
8

MUSK..
24.1 .
. 8.8
.28.1
. 3,1.3 ,
23.1
.24.9
. 7
I
. QUE..
8.3.
2.9
11.3
11.2
8.4
9.2
8 . .
RECEPTOR
S.NSC
7.4. . .
3.2
. 11.7
17.2
9.9 .
i.o.o . .
1

VT-NH
13.9 ,,
7.4.
.20.4.
, 23.9
. 16.4
. 17.3 . .
5 ...

.ADIR. .
.16.3 ..
. . . 6.0
22.8 ,
28.2 . ,
18.3 .
. 19.3 .
, . .5

PENN..
47.1. , . .
28.6 .
.70.5
79.8 ,
.56.5
60.3
6

SMOKIES
. 30.2 ...
10.1 . .
.32.1 ...
. S2..4
3.1.2 .
31.2
0

-------
                                       8-5
8-2.  (continued)
                           AVERAGE S0   CONCENTRATION



PERIOD

Spring .

Summer ....

Fall, .

Winter .
Composite
Year
Base Year
^er Cent
•Lfference



, BNDW.

1.9. .

0.4 ..

0.6

0.6

0.9 .
1.1
19.



. ALG.

.".'4.7

1.8

..2.8

1.7

2.8
3.5 .
.21 ,



MUSK. .

10..2 . .'

..4.7';

. 7.9

...3.9

6.7
8.4
.21

	 I

QUE.

. 3.6

.1.6

4.1 . .

2.2

.2.9
3.7
23

RECEPTOR

S.NSC

.3.6.

. 2.1

...5.9

4.0.

3.9.
, .5.2 .
25



VT-NH

. . 5.3

.. 3.0.

.7.2.

. . 4.0 .

. 4.9.
..6.4
23'



ADIR..

....6.8

3.8

9.9

. , . .5.2 ..

6.4 ..
. . . 8.5 . .
24



. PENN., .

12.3 .

. .8.8

1.6.8 . .

.9.0

11.7
15.3, , .
, 23 ,



SMOKIES

.13.6 . . ,

5.2

.. 1.3.7 . ,

. . 8.4.

. 10.2
.. 13.3 .
23 ,
                            TOTAL DRY DEPOSITION OF. SULFUR (kg.S.ha"!)

PERIOD.
Spring
Summer
Fall .
Winter
Composite
Year
Base Year
Per Cent
Difference

BNDW.. .
L. 0.88 .
0.77
0.46
. 0.32
2.43
. 2.44
0

ALG..
2.63
2.39
2.07
0.93
8.02
8.45
5

. MUSK.. .
, 5.02.
. 5.96 .
. 5.75
.2.57
. 19.3.0. .
20.49
. 6
I
.QUE..
1.73
. 1.96
2.34
, .0.94 ,
6.97
. 7.65
9
RECEPTOR
, S.SNC.
1.56
.2.2,5 .
..2.47
1.46
7.7,4 .
. 8.41 .
. 8 .

. .VT-NH .
2.88
. . 4.87
4.20
1.99
.. 13.94 .
.14.3.0 .
. . 3 .

. .ADIR.
3.40
4.13.
4.75
, . 2.36.
.1,4.64
. 16.13
. . 9 ..

PENN..
. 9.61
18.33
14.34
6.53. .
48.81
49.11
.1

SMOKIES
6.31
. . 6.80
.,6.6.8. ..
.4.35 .
24.14 .
.26.02
. . 7 , . .

-------
                                      8-6
8-3.  Seasonal Variation in AES-LRT Model Transfer Matrices of Per Cent of
      Total or of Annual Average (1978)

                   WET DEPOSITION OF'SULFUR AS PER CENT OF ANNUAL

PERIOD
Spring. .
Summer
Fall . .
Winter.

. BNDW , .
... 34
24
L 17
. . 25 ,

. ALG..
.38
15
30
18 . .

MUSK.
... .2.7 . .
17
33
. 22 .
I
. .QUE.
. . .25 . .
16
30
29
RECEPTOR
. S..NSC
. 18 . .
9
34
39

VT-NH
. , 20
15
, 33
33

ADIR..
20
14
33
. 33

PENN..
. 17
21,
26
36

. SMOKIES
22
.. 13 .
, 19,
. , 45 ....
                        S02 CONCENTRATION AS PER CENT OF ANNUAL. AVERAGE


Spring^
Summer
Fall
Winter, .

.BNDW..
. 147
41
78
.133

ALG.
13.5-
,'38 .
108
.120 . . .

MUSK.
104
. 38 . ,
122
136
I
QUE.
98 .
. 35
. .1.3.4. . . .
133
IECEPTOR
S.NSC
75
33
, . .119, . . ,
174

VT-NH
, 8.5 , ,
45
124
146

ADIR.
89,,
33
124
154

PENN.
83, .
. ,51 .,
. 125 .
141

SMOKIES
, 97
. . . 3.2 , . . ,
103
168
                      SO4 CONCENTRATION AS PER CENT OF ANNUAL AVERAGE

PERIOD
Spring
Sunnier .
K.
.er .

BNDW.
211
46
71
70

ALG.
170
., 65
102
62

MUSK.
153
. . 70. .
118
58
I
QUE.
126
54
143
77
RECEPTOR
S.SNC
192
55
151
103

VT-NH
109
62
147
,82

ADIR.
106. ,
59 .
154
81

PENN.
105
75
143
77

SMOKIES
133
51
134
, ,82

-------
                                 8-7
8.2.2  ASTRAP Results




     The ASTRAP model has been run using trajectory statistics



derived from January and July 1978 meterology.  The ASTRAP 1978



annual transfer matrices are based on an arithmetic mean of



the January and July matrices.  Currently, no conclusion can be



made regarding the suitability of developing an annual average



concentration or total deposition from an average of just



two months.



     The ASTRAP January and July matrices reflect not only



differences in the monthly meteorology, but also have been derived



using seasonally and diurnally varying emissions, emission release



heights, dry deposition velocities, conversion rates for S02



and 304 and eddy diffusivities.   Table 8-4 summarizes the seasonal




variation of the ASTRAP.model parameters.



     Table 8-5 presents the ASTRAP model simulations of monthly



average SC>2 concentrations and total wet sulfur depositions for



January and July 1978.  This table shows there were significant



variations in ASTRAP simulations between these 1978 winter and



summer months.  The ASTRAP monthly average SC>2 concentrations



for January at the nine receptors are on the average about 40%



higher than the July concentrations.  On the othr hand, the



wet sulfur depositions do not show a consistent difference



between January and July.  The higher S02 concentrations in



January can be attributed in part to the reduced dry deposition



velocities, and seasonal varying emissions amounts and distributions.

-------
                               8-8
Table 8-4.
Seasonal Variation in Parameter Values for the
ASTRAP Model
Parameter

V(jS02 average
(cm s~l)
V^S04 average
(cm s~l)
k average (% hr"1)
January

0.25
0.28
0.48
July

0.41
0.4.6
2.00

-------
"Cable 8-4.  January and July 1978 ASTRAP Model Estimates of SC>2 Concentrations
            and Total Wet Sulfur Depositions at the Nine Receptors
                                               Average. SQ9 Concentration, ,(M9
                                                                               _o

January
July
% Deviation
[Jan-July]
Jan

January
July
% Deviation
Boundary
Waters
0.4
0.2
50
Boundary
Waters
0.07
0.08
-14
Algoma
16.1
10.0
38
Algoma
0.90
0.60
33
Muskoka Quebec
13.3 7.0
8.7 6.7
35 4
.TOTAL WET SULFUR
Muskoka Quebec
1.25 0.71
1,08 1.37
14 -93
So.NS
2.8
1.3
54
.DEPOSITION.
So.NS
0.18
0.23
-28
Vt-NH Adirondacks
7.8 7.6
4.6 4.7
41 38
(kg.S.ha."1)
Vt-NH Adirondacks
0.88 1.66
0.99 1.61
-13 -28
W. Penn. Smokies
28.9 25.1
14.1 14.2
51 43
W. Penn. Smokies
1.93 1.52
2.55 1.55
-32 -2
                                                                                                                       00

-------
                               8-10


In addition/ the wet deposition is obviously closely correlated

with the pattern of precipitation amounts in the months under

study, and thus would not necessarily to show any correlation
                                                  •«
between the months.

     The sensitivity of ASTRAP to the mean value or diurnal

variation of its parameters has been assessed in terms of changes

in sulfur budget or changes in predicted concentration and

deposition maxima.  Since these model products are not currently

under consideration, we direct the reader to the ASTRAP Model

Profile for details.

8.2.3  Summary

     Although these preliminary results do not yet permit any

general conclusions to be drawn on the importance of seasonal

variations in concentrations and. depositions, the results indicate

that seasonal variability can be significant and should be inves-

tigated further.

-------
                              • 8-11


8.3  Model Sensitivity to Parameters

     The sensitivity of modeled sulfur concentrations and depositions

to change in values of input parameters has been investigated.
                                •»
Model parameters such as deposition velocities/ wet removal,

conversion rates of sulfur dioxide to sulfate and mixing heights

were varied, one at the time, within the range normally used for

long-range transport modeling.

     Different methods were used by the different modeling groups.

The AES-LRT model used actual meteorology and emissions and

studied the change at the targeted sensitive receptors.  The OME-

LRT and MEP-TRANS models used a single hypothetical source and

investigated the response at various distances from the source.

In the OME-LRT work the meteorology was simulated, while in the

MEP-TRANS work, the actual meteorology was used.  The results

have been expressed either as per cent deviation from the base

case value or as sensitivity indices.  The sensitivity index

gives the fractional change in computed concentration or deposition

(A) as function of factional change in parameter values (P),

defined as d(lnA)/ d(lnP).  Highlights from the evaluations are

given below while details are provided in the Model Profiles

(see Appendix 5).

-------
                               8-12


8.3.1  AES-LRT RESULTS

     The sensitivity of the modeled concentrations and deposition

amounts to changes in the values of input parameters were investiga-
                •«
ted at the sensitive receptors.  The emissions inventory, meteorology

and parameter values were the same as those used to compute the

AES-LRT transfer matrices presented in Chapter 7.

     Dry deposition of sulfur dioxide and wet deposition of sulfur

are the principle removal mechanisms for sulfur in the base case.

Thus, changes in dry deposition velocity of sulfur dioxide
                                                                2
the transformation rate of sulfur dioxide to sulfate  (k) and the

washout ratio of sulfur (W) results in the largest change in

wet deposition of sulfur.  This is clearly illustrated in Table

8-6 by the magnitude of the sensitivity indices in the rows

             W.
    2
     As an example of the changes in concentrations and deposition

at an individual site, the results  (see Model Profile for rest of

figures and tabulations) for iMuskoka are:

        Sulfur dioxide concentration is most sensitive to

        changes in dry deposition velocity of sulfur dioxide

        (index:  - 0.61) and mixing height (index:  -0.25 and 0.54),

        Sulfate concentration is most sensitive to changes in

        dry deposition velocity of sulfur dioxide (index:  - 0.40)

        and the rate of transformation of sulfur dioxide to

        sulfate (index:  0.88),

        Dry deposition of. sulfur dioxide is most sensitive to

        changes in dry deposition velocity of sulfur dioxide

        (index:  0.39) and 'mixing height (index:  -0.25 and -0.54),

-------
Table 8-6. ^Bisitivity Index - Fractional Change  in Wet Depo^Pion As a Function of
            fractional Change in Parameter Value - Annual (d In Dep/d In Parameter)

	i-i-i-i-i—•----------•--•-'----••----------••--'-•••••. '-<•:-.•.:.:----i-:~--------$.:---:.:.:--i----i----:.'^.

 Parameter               BNDW       ALG       MUSK        QUE       SNSC       VTNH       ADIR      PENN        SMOKIES
 and, flange. .... ,...'. . /,; .^/^'.'. /.^ .^\^V_/.V//^.^^                                                          .'.'.'.'.',',

 VdS02                   -0.55      -0.52     -0.50      -0.59     -0.64      -0.53      -0.53     -0.40       -0.49
 0.2-1.5 cm/s
 VdS04                   -0.14      -0.16     -0.15       -0.17     -0.28      -0.15      -0.16     -0.12       -0.19
 0.1-0.6 cm/s


 K                        0.49       0.55      0.58        0.56      0.62        0.54       0.57      0.55        0.58
 0.3-1.5%/hr


 WS02                   0.07       0.06      0.06        0.05      0.04        0.07       0.06      0.08        0.06

 3 x 103 - 3 x  104
 NA/SC-2                   0-21       0.16   .   0.18        0.15      0.11        0.18        0.16      0.28        0.21

 3 x 104 - 3 x  105


    S04                   0.34       0.39      0.41        0.37      0.46        0.36        0.39      0.42         .49

 8.5 x 104-1.7  x  106


 H (Mixing Height)        0.22       0.12      0.13        0.28      0.37        0.17        0.22     -0.11      -0.05

 (0.5 x Base H)-
 Base H
 H  (Mixing Height)        0.0       -0.14     -0.15        0.01  ,   0.04      -0.05      -0.08     -0.31      -0.24

 Base H-(2x Base  H)

-------
                       8-14


Wet deposition of sulfur dioxide is most sensitive to

changes in dry deposition velocity of sulfur dioxide

(index: -0.70) and washout ratio sulfur dioxide (linearized

index: 0.93 and 0.67),                          t

Dry deposition of sulfate is most sensitive to changes in

dry deposition velocity of sulfur dioxide (index:  -0.40'),

dry deposition velocity of sulfate (index: 0.08) and rate

of transformation of sulfur dioxide to sulfate (index:

0.88),

Wet deposition of sulfate is most sensitive to changes

in dry deposition velocity of sulfur dioxide (index:

-0.46), rate of transformation of sulfur dioxide to
                                     *
sulfate (index: 0.86) and washout ratio of sulfate (0.57),

Dry deposition of sulfur is most sensitive to changes

in dry deposition velocity of sulfur dioxide (index:

0.35) and changes in mixing height (indices: -0.25 and -0.53),

Wet deposition of sulfur is most sensitive to changes in

dry deposition velocity of sulfur dioxide (index:  -0.50),

rate of transformation of sulfur dioxide to sulfate

(index: 0.58) and washout ratio of sulfate (index: 0.41),

Total deposition of sulfur is most sensitive to changes

in mixing height (index: -0.34), the sulfate washout

ratio and transformation rate (indices: 0.15),

-------
                                .  8-15

8.3.2  QME-LRT. .R.e.su.ltS
     The sensitivity of the OME-LRT model was investigated for
an ideal case rather than at the sensitive receptors.  The idealized
source-receptor geometry, depicted in Figure 8-1, enables the testing
of the model sensitivity to the input parameters as a function of
the source-receptor orientation independent of actual meteorology.
The model parameters evaluated, apart from dry deposition velocity
of sulfur dioxide and sulfate, conversion rate of sulfur dioxide to
sulfate, mixing height and wet deposition rates, include meteorolo-
gical parameters such as the length of wet and dry periods, wind
speed and direction, and horizontal dispersion in the. along-wind
and cross-wind directions.  The sensitivity indicies for wet deposi-
tion from a source at the different receptors are shown in Table 8-7.
In summary:
Dry conversion rate (k^):
     For receptors down wind of the source the calculated wet
     deposition rate is insensitive to values of ^(.6% to 1.2%)
     except at large distances (/-*-1000 km).  The result is the
     same at cross wind receptors.  Elsewhere the calculation
     is insensitive to the value of k^.
Wet conversion rate (kw):
     At near source receptor sites the deposition is sensitive
     to values of k., greater than 5% hr'1.  Below this conver-
                   W
     sion rate there is no significant sensitivity.  The con-
     version rate must be large to show sensitivity due to the

-------
                                      8-16
                                     12 m
            18

15 *x
\
\
\
\
\
\
IK
\

\
\

\
\

\ • lOo
\
• v
	 " 	 e 	 -^
17 16
9
•
/
/
f
f
/
/
s
s »
f m
8 ,-''' ,-''''' 6
/ ,'''
' ,''


7 / ,-"*'''
f O ^4
V 	 9 	 „ (3 	 .- 	 «. 	 ... ^
1 2 3
mean wind direction
                                     source
                                                             Scale
                                                                    \	.]
                                                                    200 km
    Figure 8-1.  Idealized Source-Receptor Geometry Used for
                 OME-LRT Model Sensitivity Studies

-------
Table
OME-LRT Model Sensitivity of the Wet Sulfur      ition Factor for the Idealized
Source-Receptor Geometry Shown in Figure 8-1. (The sensitivity is presented as
the ratio of the fractional change in the deposition factor with respect to the
fractional change in the parameter).
Wet Sulfur
Deposition
Factor
(10-3 gS/m2/yr)/
(ktonne S02/yr)
PARAMETER (base case value)
KD (1% hr-1)
KW (1% hr"1)
VD (1 cm/s)
VDBAR (0.1 cm/s)
LAMDA (3.0 x lO^s"1)
LAMDAB (1.0 x lO^s"1)
MIXHT (1000 m)
FD (0.9)
FW (0.1)
UM (10 m/s)
VM (6.0 m/s)
WIND (10 m/s)
WDIR (270°)
TAU D (46 hr)
TAU W (7 hr)
Ri

5.890
(x 10~13)
0.037
0.680
-0.180
-0.003
0.560
0.052
0.169
-3.100
0.300
-1.870
-9.020
0.452
-1.240
-0.546
0.316
.R2

1.710
(x 10 i3)
0.089
2.160
-0.343
-0.009
0.384
0.063
0.336
-0.770
0.090
-0.113
-0.895
0.551
-2.040
-0.724
0.386
K3

0.580
(x 10~13)
0.165
-1.670
0.530
0.020
0.257
0.075
0.540
0.170
-0.020
-3.050
-0.865
0.648
-3.430
-0.722
0.344
W-.2

0.051
(x 10~13)
0.330
-2.200
-0.780
-0.080
0.086
0.063
0.870
0.410
-0.040
-0.094
1.090
0.998
-3.480
-0.469
0.174
R15

0.044
i (x 10 13)
0.320
-2.130
-0.770
-0.070
0.091
0.064
0.860
0.400
-0.040
0.117
0.194
-1.440
0.454
0.478
0.180
JU.8

0.049
(x 10"1)
0.290
-1.880
-0.730
-0.060
0.119
0.069
0.810
-0.390
-0.040
3.040
-0.891
-1.780
0.213
-0.531
0.221
                                                                                                        00

-------
                               . 8-18





     magnitude of other pathways for wet deposition  (e.g. wet



     deposition of SC^).  At large distances wet deposition



     is not sensitive to the wet conversion rate - again other



     pathways dominate.



Dry deposition velocity of SC>2:



     Wet deposition is sensitive to this parameter at all



     locations/ particularly at large distances.  This is due



     to the loss of sulfur by day deposition for subsequent



     wet deposition.



Dry deposition velocity of 804:



     For the range of 804 dry deposition velocities considered the



     estimate of wet deposition does not depend on this parameter



     (insignificant pathway).



Wet deposition rate of SC^:



     Wet deposition is very sensitive to this parameter downwind



     and near the source.  This pathway near the source can



     result in greater than 80% of the wet deposited sulfur to



     be SC>2.  Far from the source wet deposition is less



     sensitive to this parameter as other pathways to wet deposi-



     tion become important and SC>2 has been depleted or converted.



Wet deposition rate of 804:



     This pathway for wet sulfur deposition is marginally sensi-



     tive to the value of deposition rate as the total amount



     of wet 804 available for scavenging is controlled by other



     model parameters (e.g. kw

-------
                                8-19

Mixed layer height:
     Due to the assumption of vertical homogeneity an increased
     mixed layer height reduces the ground level concentration
     thereby reducing the loss by dry deposition while the wet
     scavenging rate remains constant.  Therefore at all loca-
     tions the wet deposition rate increases as the mixed height
     increases.
Fraction of time (Eulerian) of dry (f^) or wet  (fw) conditions:
     The estimated wet deposition is sensitive to these para-
     meters close to the source downwind.  These locations
     are near enough to the source that the nature of the
     pollutant (wet or dry) emitted is not 'forgotten1.  (i.e.
     travel time is less than the time taken to lose the particles
     identity  ("Crw or tr^) •  For these regions, if we increase
     the fraction emitted wet then wet deposition increases propor-
     tionally.
Lagrangian duration of wet (T^) and dry (Tr^) periods:
     The model estimate of wet deposition is sensitive to these
     parameters for all source-receptor orientations as the
     parameter control rate of exchange between wet and dry
     particles.
Horizontal/ along wind dispersion ( 6~u):
     The model estimate of wet deposition is sensitive to this
     parameter as it controls the probability of the emission
     reaching a receptor.  Least sensitive are the directly down-
     wind sites which do not rely on dispersion to transport

-------
                                8-20






     the emissions.  Conversely upwind sites depend solely on



     dispersion and are the most dependent on this parameter.



Horizontal, cross-wind dispersion ( £~~v):



     All receptor sites are sensitive to this parameter.  As




     for   (TU' this parameter controls the value of the probability



     of arrival at the site. Sites least affected are those downwind



     near source sites in the cross-wind direction that are



     due to depletion of the pollutants by cross-wind spread.



Mean wind speed:



     All receptor sites are sensitive to this parameter.  Wet



     deposition at a site will increase or decrease with increased



     wind speed depending upon whether increased wind speed will



     move the pollutant toward or away from-the receptor more



     quickly.



Wind direction:



     All downwind sites are sensitive to the wind direction;



     however, for the range chosen (255° - 285°) directly up-



     wind sites do not show any sensitivity.  Deposition is



     maximum for those sites directly down wind.



Additional results are provided in the model profile.



8.3.3 M.E.P TRANS Results



     The sensitivity analysis was carried out for a single source



by annual runs with the range of parameter values indicated in



Table 8-8, varying only one parameter at a time.  The results for

-------
           Table 8-8.  Range of Parameter Variation for the MEP-TRANS Model Sensitivity Analysis
Parameter
                Sulfur                                       Nitrogen
     Summer                  Winter               Summer                     Winter

Low    Std.    High    Low    Std.   High    Low    Std.    High     Low    Std.   High
Primary Deposition
  Velocity (cm/s)
0.1    0.8     2.0     0.1    0.3    2.0      0.1     0.5      2.0      0.1    0.3    1.0
Secondary Deposition
  Velocity (cm/s)
0.1    0.4     2.0     0.1    0.3    1.0      0.1     0.5      2.0     0.1    0.3    1.0
Transformation Rate
  (% hr -1)
Primary Washout Rate
  (hr -1)
1.0    2.0     4.0     0.5    1.0    2.0      3.0     5.0     10.0      1.5    2.0    4.0
               Not Varied
0.01   0.04    0.1     0.01   0.04   0.1
                                                                                            T
                                                                                            |SJ
Secondary Washout Rate     0.04   0.5     0.8     0.04   0.5    0.8      0.04    0.5      0.8      0.04   0.5    0.8
  (hr -1)

-------
                               8-22






the sensitivity of 804 wet deposition as an example of the analysis



performed for the parameters are presented in Table 8-9.



     In Table 8-8, the average 864 deposition is the average of



all receptors within each of the distance ranges indicated.  Thus,



the average deposition within 100 km from the source is 42 mg



S/m^/yr for the standard choice of the parameters.  The percentage



change is then the average over the receptors within the given



distance range.  Thus, choosing the low value for transformation



rate reduces the wet deposition in the near range by 47%, in



the mid-range the reduction is only 34% and in the far range it



is only 23%.  The high transformation rate increases the washout in



the short-range by 80%, but has little effect in the long range.



This reflects the fact that removal of 804 by precipitation is more



efficient than SC>2 removal.  Reducing the deposition velocities,



particularly the S02 deposition velocity, increases the 804 wet



deposition; in this case the effect being cumulative with distance.



The effect of 804 washout rate on 864 wet deposition would seem to



be contradictory in that a reversal of the expected effect is



observed.  Lowering the washout rate does lower the wet deposition



in the short range, but increases it in the far range, since the low



washout rate allows accumulation of sulfate which can be washed out



at longer range.

-------
Table 8-9.  Sensitivity of Wet 804 Deposition to Variations in MEP-TRANS Model Parameters
Distance
From Source
. , , .(km) 	
0 -
100 -
200 -
500 -
700 -
1000 -
>
100
200
500
700
1000
1500
1500
Average 804
Wet Dep. (mg S m~2)
. Standard, Parameter, .
42.
20.
6.
1.
0.
0.
0.
3
2
9
9
7
2
1
Transformation
Rate
	 .Low 	 , fligh. . .
-47
-45
-39
-34
-30
-23
-23
+80
+72
+47
+30
+19
-2
0
Deposition Velocity
so2 so4
.... .Low 	 High . , , .Low. . . , High
+18
+27
+55
+77
+98
+140
+143
-26
-32
-46
-54
-60
-72
-75
+5
+7
+13
+21
+20
+22
+21
-16
-20
-31
-43
-40
-40
-39
Washout Rate
of SO,
. . LOW. 	 ttigh.
-52
-38
+28
+87
+173
+433
+407
+3
-2
-11
-21
-22
-21
-25
                                                                                                                     00

-------
                              •  8-24



     This example points out the complex dependence on the several



removal and transformation processes within the model, and shows



that one cannot consider the various parameters as being truly



independent.



     For further results see the Model Profil-e.



8.3.4  Other Model Results



     Sensitivity analyses for the RCDM, ENAMAP, CAPITA Monte Carlo



and UMACID models will be provided in an addendum to the Modeling



Subgroup Report by September 1981.

-------
                          Chapter 9
                 PRELIMINARY SOURCE-RECEPTOR
              RELATIONSHIPS FOR NITROGEN OXIDES
9.1  Introduction


     At the present time the basic principles of the atmos-

pheric chemistry of oxides of nitrogen/ at least with respect

to the conversion from primary pollutant (NO) to the ultimate

secondary pollutant (HN03) are reasonably well characterized

via experimental smog chamber studies.  The evaluation and

verification of these processes under atmospheric conditions

are far less established.  The role of ozone and peroxy-radicals

in converting NO to N02 is well established/ and the subsequent

formation of HN03 is known to be due to the reaction between

OH radicals and N02.  It is also well documented that/ con-

currently with HN03 formation, PAN production will occur pro-

vided that sufficient acetaldehyde is available.  Although

actual observations concerning PAN in the atmosphere are

limited in number, they do suggest that under most conditions

when NO is emitted from a source, such as transportation

activities, enough hydrocarbons are indeed co-emitted to

cause PAN formation.  Finally, from our present understanding

of these processes it appears that most of the reactions occur

exclusively in the gas phase and hence/ contrary to the SOX

chemistry, one is not burdened by the well-known complications

caused by liquid phase chemistry in the atmosphere.

-------
                             9-2





     It is at present difficult to model the NOX chemistry



with the same confidence as the SOX chemistry for the



following two reasons:



     Firstly, the SOX chemistry is simplified to the ultimate



limit with the linear parameterized oxidation reaction S02—>



sulphate.  From a chemical viewpoint it can be argued that



this is feasable since the conversion is relatively slow



(S02 has a chemical lifetime of a few days) and hence



averaging of diurnal variations is acceptable.



     The common use of a timestep of more than one hour in



model runs is then also reasonable.  On the other hand, the



NOX oxidation chemistry takes place at faster rates, (the



oxidation of NO—» N02 within the timespan of a few hours,



while the chemical lifetime of N02 is less than one day)



and consequently the use of diurnally averaged rates as well



as the application of timesteps of a few hours is questionable.



     In connection with this matter, it can be stated that



the SOX oxidation process has a limited influence on the



concentration of oxidizing species (hence first order chemistry



can be assumed), while the NOX oxidation process has a



strong feedback effect on the levels of oxidizing species.

-------
                             9-3





     The second major problem rests with the evaluation



of the modeling efforts.  The simplified SOX chemistry in



the current LRTAP models can at least be defended with the



qualified statement that comparison with data from actual



observations in the field shows that the modelled data are



not very divergent.



     Such an evaluation of potential NOX chemistry modeling



is not possible at the present time since the data base on



field observations is very poor indeed.  The networks that



have been in operation to date have produced data on wet



deposition of nitrate, but it is highly doubtful whether



they represent actual levels in the rain due to factors such



as liquid/gas exchange of the nitrate ion, and in some cases



biological activity in the samples.  They produce no data on



levels of particulate nitrate or gaseous HNC>3 in the U.S.



The data from the APN and Ontario Hydro networks have unknown



(but potentially very large) errors.  Finally, nowhere do



the networks produce data on concurrent levels of NOX or



PAN.  Hence it is not even.possible to start evaluating



whether the complications that are anticipated from our



knowledge of the NC>2 chemistry are indeed serious enough to



make NOX modeling with the current LRTAP models impossible



or not.  In addition, N02 measurements (when available at

-------
                             9-4  •

rural monitoring sites) are typically below the lowest detect-
able limits of currently available automated monitoring
equipment.
     There is still another matter which causes NOX modeling
to be almost pure speculation at the present time.  As can
be grasped from the review by Lusis and Shenfeld  (1981) on
the seasonal effects on chemistry parameterization, there
is at least a limited amount of information available to
make guestimates on how to treat the deposition processes
of S02 and sulphate in the models.  Similar information is
virtually nonexistant for the nitrogen oxide species, and
one is forced even more to use chemical and physical intuition
when parameterizing loss processes from the atmosphere.
^ • 2  .P.a.r.ain.e.t e r i z a.t,i.on
     Three preliminary efforts to model the chemistry of nitro-
gen oxides with LRTAP models appear to have been made.  A
"minimal" chemistry adequate according to Jeffries  (1979),
based on smog chamber studies, is shown in Figure 9-1.
Obviously LRTAP models cannot now incorparate such  chemical
detail.  The CAPITA Monte Carlo simulation model  (Patterson
et al.  1981) has been so far the most extensive in its
chemical considerations by applying a scheme shown  in Figure
9-2.  The CAPITA model relies upon diagnostic determination

-------
 Figure 9-1. Gas Phase Nitrogen Oxides Chemistry
        (Jefferies, 1979).
  (RQNQ)
    .   <-£
mo NO
                   RONQj
                                                     U1

-------
                              9-6
                              HN03
                              PAN -
                                            •HN03 DrY Deposition
  HNC>3 Wet  Deposition
^»-PAN  Deposition
                              N02  Dry  Deposition

                              'NC>2  Wet  Deposition
Figure 9-2. The Kinetic Scheme  for  NOX  Modeling Used By
            the CAPITA Model.

-------
                             9-7





of rate parameters using ambient concentration and source



inventory data.  Since ambient data are virtually nonexistent,



the applied rates were selected with consideration of the



overall chemistry and numerous other facts and hints that are



not documented (see Table 9-1)/ but very little information



was provided to support the quantitative values used.  This



in mind, the following results were obtained from the different



models.



     The MEP-TRANS model further simplifies the chemistry



by using the the first order reaction N02—-WJ03 as the only



chemical reaction.  Rate constants and deposition parameters



are presented without discussion on the choice of parameters



(see Table 9-2).   The AES model also simplifies the chemistry



to the single reaction NC>2—>NC>3 (Bottenheim, 1981).



Here it is argued that NO—»NC>2 conversion is too



fast for the timeframe of the model to warrant separate



consideration, while PAN chemistry is deleted for lack of



any field data.  Conversion rate is derived from the theoretical



rate constant ratio kHO+NO AHO+SO  / assuming the majority



of SC>2 oxidation to be due to gas phase chemistry which



results in a conversion rate of 10%h~^.  Deposition parameters



are from the sparse literature, see Table 9-3.  For source



emissions all three models have used the same inventories,



i.e. those reported by Voldner et al., (1980) for Canada,



and Clark (1980)  for the eastern U.S., which are presumably



good for 1978.

-------
                              9-8
 Table 9-1.    NO  Kinetic Rate Constants (h"1) Used By
               the CAPITA Model.
TRANSFORMATION
                                                  DEPOSITION




                                               DN02   DN03   DPAN

July


October


January


April

Fast

Slow
Fast

Slow
Fast

Slow
Fast

Slow
0.

0.
0.

6.
0.

0.
0.

0.
5

1
4

08
25

05
4

08
0.

0.
0.

0.
0.

0.
0.

0.
04

02
03

015
015

008
03

015
0.

0.
0.

0.
0.

0.
0.

0.
04

02
03

006
02

004
03

006
0.

0.
0.

0.
0.

0.
0.

0.
02

01
015

008
01

005
015

008
0.06

0.03
0.05

0.025
0.03

0.05
0.05

0.025
0.02

0.02
0.015

0.015
0.01

0.01
0.015

0.015
        Fast    0.038     0.029




YEAR AVERAGE




        Slow    0.78      0.015
                 0.03
0.015  0.048  0.015
                 0.009    0.008  0.024  0.015

-------
                             9-9
Table 9-2.  Parameter Choice For the 1978 MEP-TRANS Model
            Simulations of NOX.
Parameter                            Nitrogen

                                 Summer    Winter

Primary Deposition                 0.5      0.3
Velocity (cms"-'-)

Secondary Deposition
Velocity (cms"1)                   0.1      0.1

Transformation
Rate (% hr"1)                      5.0      2.0

Primary Washout
Rate (hr"1)                        0.04     0.04

Secondary Washout
Rate (hr"1)                        0.3      0.3

Mixing Height (m)                  750      500
     Table 9-3.  Deposition Parameters For NOX Chemistry
                 Used in the AES-LRT Model.
'
                        cms"1             W
N02                 0.2a              1/4 x WSQ b

HN03                1.0C              1/2 x WHS0
a Average from Bottger et al., 1978

b Beilke, 1970

c Estimate based on no surface resistance

d Estimated from work of Levine and Schwartz (1981)

W Scavenging ratio (dimensionless)

-------
                             9-1-0
9.3  Mode.l.ing/ re.suIts



     It will be clear by now that with so little data  avail-



able to anchor the parameterization, any modeling runs to



date cannot be considered to yield more than some educated



speculation of the role of nitrogen oxides in LRTAP.  With



     CAPITA, mp.de.1



     The CAPITA model has not been run for specific  source-



receptor combinations, but overall budgeting has been



attempted for the total Eastern North American continent.



It is predicted that on a yearly average basis 50-65% of the



total emitted nitrogen is transported out of the region con-



sidered (i.e.  Eastern U.S. and Canada combined), the  largest



export occurring during the winter (least HNC>3 formation).



With the kinetics employed, 17-20% of the emission is exported



as HNC>3, 10-16% in the U.S. and 4% in the Canadian region



considered.  As far as transboundary flow is concerned, the



model predicts that of the total Eastern North American



emissions, 11-12% is transported from the U.S. to Canada, of



which about 8% is deposited (the remaining 3-4% is transported



further beyond the region considered).  In contrast, 2% is



exported from Canada to the U.S. of which 1% is deposited.



Similar figures are predicted for net export of nitrogen and



HNC>3 deposition.  In summary, then, transboundary flow of



nitrogen from the U.S. to Canada exceeds the Canadian  nitrogen



export by a wide margin, according to the model.

-------
                             9-11





     HEP-TRANS model



     The MEP model has been run to produce annual average



concentration and deposition fields for 1978.  High concen-



trations for NC>2 are predicted for the Detroit and New York



areas/ while N03 (as HNC>3 or as a component of TSP) peaks



in the New York area only.  Total N deposition is predicted



in excess of 1.5 g m~2y-l in the lower Great Lakes region.



The MEP model has also been used to produce nitrogen transfer



matrices.  The transfer matrix for total N is presented in



Table 9-4.



     It is interesting to note that even if data on nitrogen



oxides are fragmentary and uncertain at this time that with



the simple nitrogen chemistry used, the wet NC>3 to 804



deposition in Southern Ontario and Quebec are in a ratio of



1 to 2, in agreement with a number of experimental studies



carried out in these areas.



     the'AES-model



     The AES model has been run only with the total emission



inventory/ to produce monthly and yearly average airborne and



deposition data for 1978 at some of the sensitive receptor



areas.  These results are presented in Figures 9-3 to 9-5.



The correlation between observed and predicted data/ Table



9-5, is generally poor/ although it is remarkable that the



best correlation is obtained for the two receptor areas



(Pennsylvania and Muskoka) which are closest to large source



areas.

-------
                               Table 9-4.  Normalized Transfer l^Bix  For Total Nitrogen From the MEP-TRANS
                                           Model  (kg.N.ha."1 yr.^- per Tg.N.yr."1) ..
                              _...._........,.,..., ...^..'. '..' V  ' '..'.' ' '..' '.'..'.' .RECEPTORS;.  '_'_/ '.'.'.'.'.'.' ' '_.'..'.'.' '//.'..'/./_.'_,'_.'.,'. '_,V._.'.
                           #1     #2        #3      #4       #5      #6       #7    #8      #9       #10       #11
REGIONS     Q:NO2       , 3..^ Wft.tgrs. , Alg;.. . MusK ., ^ Jjuie .^ ^ _. JS. .JSUJSc.., Jfl1 . NJH .,^ Adirj.^ _gejryv gjngKiea FJ.Qr.idA
         (x  TG N/YR)

 # 1 Michigan    0.469 |  0.15 |  2.37  | 2.38  |  0.35 |   0.16  |  0.47  |  0.63 |  1.44 |   0.27  |   0.03 |   0.07
 # 2 111. ,Ind.    0.724 |  0.09 |  0.64  | 0.58  |  0.13 |   0.04  |  0.18  |  0.23 |  0.39 |   0.98  |   0.17 |   0.57


 # 3 Ohio        0.606 |  0.02 |  0.25  | 1.97  j  0.32 |   0.16  |  0.48  |  0.74 |  3.96 |   0.70  |   0.06 |   0.04


 # 4 Penn        0.303 |  0.01 |  0.12  | 1.11  |  0.66 |   0.46  |  1.40  |  2.33 |14.83 |   0.17  |   0.04 |   0.00


 # 5 NY to Maine  0.906 |  0.01   0.06  | 0.63  |  1.63 |   1.68    4.16  |  4.57   1.02 |   0.02     0.03 |   0.00
 # 6 Kent^Tenn    0.336 |  0.02 |  0.08  | 0.13  |  0.03 |   0.03  | 0.06  |  0.09 |  0.35 |  11.72  |   0.82 |   0.17


 # 7 N.Va to NC   0.541 |  0.01 |  0.04  | 0.21  |  0.10 |   0.09  | 0.24  |  0.34 |  1.93 |   3.56  |   0.30 |   0.01


 # 8 Rest of      1.189 |  1.79 |  0.67  | 0.25  |  0.05 |   0.03  | 0.06  |  0.09 |  0.13 |   0.50  |   1.04 |   1.05
      East, us          ..._._,......_....	_.^:^_._^^:__,_,^^^_,_.:\_^\,^^^_,_.  .^^_._,, ^_.^._._^_l_.',_^^_.,^^_.^, ._,'.\.:,,_

 # 9 Ontario      0.150 |  0.06 |  2.30  |11.76  |  1.44 |   0.44  | 1.93  |  3.55 |  3.28 |   0.16  |   0.01 |   0.01


 #10 Quebec       0.131 |  0.02 |  0.20  | 1.50  |  9.61 |   1.62  |10.97  |18.55 |  0.28 |   0.02  |   0.00 |   0.00


 #11 Atlantic     0.055 |  0.00 |  0.02  | 0.10  |  0.56 |   4.69  | 0.31  |  0.16 |  0.10 |   0.00  |   0.00 |   0.00
     Provinces          .^^^^^^^^^^^^^^_._^^^_.^^^^^^^_!_!_.^^^_._..^_,_._._.^^^^^^^^^^_,^^^^^_l_!_l_l_'.^_._._'.^_

 #12 Sudbury      0.003 |  0.06 |  3.80  | 7.06  |  1.57 |   0.27  | 1.75  |  3.50 |  1.33 |   0.07  |   0.00 |   0.01


 #13 West         0.061 |  3.61 |  0.39  | 0.08  |  0.03 |   0.00  | 0.03  |  0.04 |  0.01 |   0.04  |   0.01 |   0.08
      Canada            .,	,	'.'...'.'...'. '.'.'._._,_._.^.._._.....^^_«.- .^^^^-^-^-.' • • > • • • • -^-

-------
                             9-13
        Table 9-5.   Correlations Between Observed and
                    Predicted N03 from the AES-LRT Model,
(ATIKOKAN,  CANSAP):   r =  1.5   + 3.2


(WAWA,  CANSAP):   r =  2.3   -  .27


(PETERBOROUGH,  CANSAP):   r =  .89  +  .91

(PETERBOROUGH,  HYDRO):  r=  -.11  +1.07

(QUEBEC,  CANSAP):   r =  7.8   - 1.08


(SHELBOURNE,  CANSAP):  r =  1.26  -  .18

(REST OF  EASTERN U.S., MAP3S):   r =  1.20  +  .07

(PENNSYLVANIA,  MAP3S):  r =   .11  +  .36
(BOUNDARY WATERS,  CALCULATED):   r=  .09 (n = 11)


(ALGOMA,  CALCULATED):   r = -.14 (n = 8)

(MUSKOKA, CALCULATED):   r =  .43 (n = 10)

(MUSKOKA, CALCULATED):   r =  .48 (n = 12)

(QUEBEC,  CALCULATED):   r = -.35 (n = 8)

(NOVA SCOTIA,  CALCULATED):  r = -.19 (n = 11)

(REST OF  EASTERN U.S.,  CALCULATED):  r =  .25 (n = 11)


(PENNSYLVANIA,  CALCULATED):  r =  .68 (n = 12)

-------
                              9-14

9.4  C.o.n.cius ipjh4 .a,nd. R.e.c.orhpi.e.hd.ait.i.on.s
     Modeling of nitrogen oxides chemistry  is  in  its  infancy.
In fact the data discussed in the  last  section must be  viewed
more as speculative exercises which require  extensive further
research before resemblance with the  real world can be  assumed.
As stated previously:  (1) the time and  space scales of  present
LRT models are not sufficiently resolved to  adequately  treat
the complex non-linear transformations  associated with  nitrogen
oxide chemistry, and (2) the  alternate  evaluation method
which is heavily used  for the SOX  chemistry, and  which  consists
of comparison with observational data,  is impossible  for  NOX
chemistry due to the lack of  data.  The best statement  that
can be made at the present time is that our  scientific  knowledge
is inadequate to quantify the NOX-HN03  problem.  Regional
scale models are urgently needed to evaluate the  time-space
scale for the nitrogen oxides chemistry.  More reliable
observational data are especially  needed.   For policy considera-
tions, no quantitative recommendations  from  the LRTAP models
can be extracted at the present time.   In a  qualitative
sense it appears NOX-HNC>3 transport should  be  more limited in
average distance travelled than SOX transport, and hence
should be more of a sub-regional problem.
     With respect to transboundary flow one  can point to  the
large source areas and prevailing  atmospheric  flows,  and

-------
                             9-15
suspect that sources in the lower Great Lakes area will have



considerably more impact on Canada, but that the other



large source area (The Washington-Boston corridor) will



probably have a lesser impact because of the greater distance



from Atlantic Canada.  Conversely the impact of Canadian



sources on the U.S.  should be relatively small in view of



their smaller size,  and should mainly come from the southern



Ontario region.

-------
                          Chapter  10


          PRELIMINARY/S.O.U.RCE-RE.CE.PTOR .RELATI.ONSH.IPS
                     FOR PRIMARY SULFATES
10.1  ,In,t.r.od,uc.t.i.o.n

     In the Phase I Interim Report, Work  Group  3A  recommended

that Work Group 2 consider assessing  the  relative  contribution

to acid deposition of primary sulfate emissions from oil-fired

and coal-fired combustion sources compared with that from

secondarily formed sulfates.  The Work Group  decided to

first review the previous primary sulfate modeling efforts

as to their adequacy for use in Phase II.  The  two principal

primary sulfate modeling efforts that were reviewed were

those by Dr. J. Shannon of Argonne National Laboratory  for

the Environmental Sciences Research Laboratory  of  the U.S.

Environmental Protection Agency and by PEDCO,  Inc. for  the

Morgantown Energy Technology Center of the U.S.  Department

of Energy.  In addition, Work Group 2, in cooperation with

Work Group 3B, reviewed the three principal primary sulfate

emission inventories for the eastern  U.S., namely, the  EPRI

(SURE Phase II), the PEDCO, and the EPA  (Homolya).  Work

Groups 2 and 3B concluded there were  significant differences

between the three principal primary sulfate inventories for

the eastern U.S.  which would have to be  reconciled in  Phase III

A primary sulfate emission inventory  for  Canadian  sources

would also be prepared in Phase III.

     The Work Group found some major  deficiencies  in the  PEDCO

modeling (Szabo, et al.,1981) of primary  sulfate emissions

-------
                             10-2
and, therefore decided to use the Shannon-Argonne National
Laboratory results in the Phase II report  (Shannon, 1981).
Additional modeling work using the participating models
and improved data bases especially for primary sulfate
emissions is planned in Phase III.
     Some of the major deficiencies in the PEDCO modeling of
primary sulfate emissions that the Work Group found were:
     (1)  the use of an incomplete SC>2 emissions inventory —
          only the top 50 SC>2 emitters representing less than
          50% of the total eastern U.S. emissions were used
          for the distant sources and SC>2 emissions for New
          York, eastern Pennsylvania, New Jersey, and New
          England were used for local sources;
     (2)  the use of a very high sulfur dioxide to sulfate
          conversion rate due to catalytic oxidation processes
          that may only apply to some sources at night;
     (3)  the use of a very simplistic box model which is not
          appropriate for determining either short range or
          long range impacts; and
     (4)  the use of modeling results which even if accepted
          as correct, do not provide conclusive evidence that
          local sources are the major contributor to acidity
          in sensitive areas like the Adirondacks.

-------
                             10-3
10.2  Em.i.s.s.ioAS



     Previous regional modeling efforts have assumed primary



sulfate emission factors typically of about 2%; i.e., 2% of



the pollutant sulfur emitted in the form of sulfate and the



other 98% in the form of SC>2.  Since the average rate of



chemical transformation of 862 to sulfate in the free



atmosphere has been assumed to be about 1% hr~l, shortly



after emission regional model simulations show that secondary



sulfate from atmospheric transformation dominates primary



sulfate (e.g., Shannon, 1981).



     However, some recent field studies involving measurements



of primary sulfate and S02 emissions from large package



boilers for apartment complexes or hospitals in the New York



City area during winter have indicated a primary sulfate



emission factor of 13.4% (Homolya and Lambert, 1980).  The



boilers burned residual fuel oil of 0.3% sulfur content.



While the total sulfur oxide emissions from such low-sulfur



fuel sources are only a fraction of total sulfur oxide emissions



in the eastern U.S., the significantly larger primary sulfate



emission factor calls for a more detailed investigation of



the relative importance of primary and secondary sulfate in



areas with those package boilers.



     The sulfur oxide emission inventory, compiled as part of the



Sulfur Regional Experiment (SURE) study, is classified by fuel



type (coal, residual oil, distillate oil, and miscellaneous)



for point sources larger than 10,000 tons SOX per years (source)

-------
                             10-4





or 1000 tons SOX per year (stack), and by source type



(residential, commercial, industrial, transportation, and



small point sources) for area sources on a grid with spatial



resolution 80 km.  This makes the inventory quite convenient



for an examination of the relative contributions of primary



and secondary sulfate.



     In Shannon's (1981) analysis the SURE emission inventory



was classified into five subsets for both winter and summer.



Miscellaneous large point sources were combined with large



point sources burning distillate oil, and the various area



source categories were combined, but split geographically



into northeastern and southwestern regions.  The sulfate



emission factors were assumed to be independent of season,



but the emission totals, particularly for the area sources,



varied considerably by season, and thus separate model simula-



tions were run for summer and winter.  Table 10-1 lists the



assumed emission factors and the emission rate for each



subset of sources.  Here, the northeastern region is defined



as being north of 36 degrees latitude and east of 78 degrees



longitude.  Weighting of the primary sulfate emission factors



by the emission total of each category produces an average



primary sulfate emission factor  of 3.0% in summer and 3.6%



in winter.  The seasonal change in the average emission



factor is due mostly to the large increase in northeastern



area sources which have the highest sulfate emission factor.

-------
                                10-5
       Table 10-1:  Sulfate Emission Factors and Sulfur Oxide
                    Emission Rates.-
     SOURCE CATEGORY             EMISSION RATE           SULFATE  EMISSION
                             (equivalent tons S02/day)        FACTOR

                                summer    winter

Coal point sources              42,000    40,000               1.5%

Residual oil point sources       4,t)00     4,000               7.0%

Distillate oil point sources     6,800     6,800               3.0%

Northeastern area sources        4,500     4,500             13.4%

Southwestern area sources      . .19.,.3.0,0    .27,,.3.0.0               3.0%

                                76,600    88,400



   10.3  Aff,TRAP Model. Re.s.ul.ts

        Previous simulations with the Advanced Statistical Trajec-

   tory Regional Air Pollution (ASTRAP) model (Shannon,  1980)

   had attempted to account  for some of the variation  in primary

   sulfate emission factors  by adjusting the factor by emission

   level (layer).  The adjustment was not the same in  all

   simulations, but typical  sulfate emission factors ranged from

   3.0% in the lowest layer  to 1.8% in layers above 400 m or so.

   The weighted average factor was thus a bit more than  2%,

   still consistent with the typical sulfate emission  factors

   in regional models mentioned above.  In order to investigate

   the principal sources of  anthropogenic sulfate in greater

   detail, the ASTRAP model  was modified to calculate  the norma-

   lized vertical profiles of both primary and secondary sulfate,

-------
                             10-6





rather than a single normalized vertical profile of total



sulfate as in previous work.  Primary sulfate factors were



included as a parameterization in the concentration and



deposition subprogram of ASTRAP, which was exercised



separately for each of the emission subsets in Table 10-1.



The same January and July 1985 meteorological data were used



as in previous ASTRAP applications.



     The change in ASTRAP formulations described above



produce only minor changes in the simulations of average S02



concentration and of dry and wet deposition of total sulfur,



since the wet removal parameterization in ASTRAP is indepen-



dent of sulfur species, the dry removal parameterization is



quantitatively similar for each species, and because the



large relative increase (50-75%) in the average sulfate



emission factor leads to only a 1-2% relative reduction in



the S02 emission factor.  Therefore, our detailed examination



focuses on sulfate concentrations.



     ASTRAP simulations of the average surface concentrations



of primary and secondary sulfate resulting from all eastern



North America anthropogenic sources in summer and in winter



are shown in Figures 10-1 and 10-2 and 10-3 and 10-4, respec-



tively.  Viewed from the broad perspective of the eastern



half of North America, secondary sulfate still dominates



primary sulfate.  However, the. winter sulfate patterns in



New England and the eastern half of the Middle Atlantic states

-------
                             10-7
Figure 10-1,
Isopleths of Summer (July - August) Primary
Sulfate Concentrations (ug m~3)  Simulated
by the ASTRAP Model (Maximum Concentration
3.7 ug m~3).

-------
                             10-8
                                          V   \
Figure 10-2.
Isopleths of Summer (July - August) Secondary
Sulfate Concentrations (ug m~3 )  Simulated
the ASTRAP Model (Maximum Concentration
14.9 ug m~3).

-------
                            10-9
Figure 10-3.
Isopleths of V^inter (January - February) Primary
Sulfate Concentrations (ug m~3) Simulated by the
ASTRAP Model (Maximum Concentration 7.7 ug m~3).

-------
                            10-10
Figure 10-4.
Isopleths of Winter (January - February) Secondary
Sulfate Concentrations (ug m~3)  Simulated by the
ASTRAP Model (Maximum Concentration 6.3 ug m~3).

-------
                             10-11





show that primary sulfate concentrations are as large as secondary



sulfate concentrations in the Washington-Boston urban region.



In fact, in the New York City area, primary sulfate concentra-



tions are larger than secondary sulfate during winter.  If



the area source resolution were improved below the 80 km



squares of the SURE emission inventory, gradients of sulfate



concentration between urban and rural areas in the Northeast



produced by superimposing the broad regional patterns of



secondary sulfate and the relatively spiked urban patterns



of primary sulfate would probably be even steeper.



     Some additional insight to the sulfate patterns may be



gained from examination of separate simulations of point and



area sources.  The large point sources, upon which most



regulatory emphasis rests, are generally elevated, some well



above the heights reached by nocturnal inversions, and are



concentrated in the Ohio River Valley.  Because of the eleva-



tions of the point sources, the plumes are decoupled from



surface removal processes during most nights so secondary



sulfate can increase before the sulfate plume contacts the



surface the next day.  In addition, the average sulfate



emission factor for point sources is quite low, because of



the preponderance of coal combustion.  For these reasons,



one would expect the dominance of secondary over primary



sulfate from point sources.  Indeed, the model results for



the summer and winter cases, respectively, shows that primary

-------
                            10-12





sulfate from large point sources is only a small fraction as



great as secondary sulfate/ on a regional basis.  The disparity



is greater in summer than in winter, because of the higher



rate of S02  conversion during warm, humid weather.



     When one examines the simulations of emissions from area



sources, however, a different pattern appears.  First, the



area sources actually consist of a very large number of small •



point sources which are normally emitted near the surface.



The ASTRAP model simulations assume that area emissions are



initially evenly dispersed through the bottom layer (0-100



m).  Thus, the decoupling mechanism is unimportant and the



effect of primary sulfate is felt at the surface immediately.



Second, the area sources are more important during the winter



season, when space heating is required, and that is the



season when atmospheric transformation of SC>2 is least



effective.  Examination of the sulfate concentrations from



area sources for summer and winter shows that in the northeastern



urbanized corridor, primary sulfate concentrations are about



half as large as secondary sulfate concentrations during the



summer, but more than twice as large as secondary sulfate



during the winter.



     Mention should be made here that the ASTRAP model simulates



only anthropogenic sulfate from the area east of the Mississippi



River.  The flux of anthropogenic sulfate from the west is



not included in ASTRAP so the gradient of sulfate in the west

-------
                            10-13





part is steeper than might be expected in nature.  In addition,



there was no attempt to add a background sulfate concentration



for biogenic sulfate.  In addition, the SURE inventory is also



known to underestimate eastern Canadian S02 emissions and



caution is required in the interpretation of simulation results



in areas impacted by eastern Canadian sources.  A final and



most important caveat is that the accuracy of the above



analysis is dependent upon the accuracy of the assumptions



about sulfate emission factors in Table 10-1.



     Source/receptor transfer matrices for primary and



secondary sulfate concentrations are shown in Tables 10-1 and



10-2.  It should be noted that the emission data (Table 6-1)



and meteorological fields used in generating the transfer matrices



are not identical with the input data for the simulations



shown previously in Figures 10-1 through 10-4.

-------
                                          10-14

Table 10-2.  Transfer Matrix for January 1978 Sulfur Dioxide and Sulfate
             Concentrations (ug/m3) from the ASTRAP Model Using the Phase III
             State/Provinces SC>2 Emission Inventory and Primary Sulfate Emission
             Factors in Table 10-1.
S02
Spurce Regions
1
2
3
4
5
6
7
8
9
10
11
Total
Primary 864
Source Regions
1
2
3
4
5
6
7
8
9
10
11
Total
1
0.01
0.03
0.01
o.oo •
0.00
0.00
0.00
0.11
0.36
0.01
0.00
0753

0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.01
0.02
0.00
0..00
"OT
2
0.04
0.10
0.03
' 0.01
0.00
0.01
0.00
0.18
1.56
0.14
,0.00
"2J07

0.00
0.01
0.00
0.00
0.00
0.00
0.00
0.01
0.05
0.01
0.00
'OF
3
0.27
0.35
0.11
0.02
0.00
0.05
0.00
0.47
3.60
0.41
0.00
5.28

0.02
0.02
0.01
0.00
0.00
0.00
0.00
0.03
0.10
0.01
0.00
"07211
Receptor Regions
456
0.14 0.06 0.35
0.23
0.18
0.14
0.14
0.06
0.04
0.15
1.21
6.92
0.03
9.23

0.01
0.02
0.01
0.02
0.03
0.00
0.00
0.01
0.05
0.79
.0...01
0.10
0.09
0.16
0.34
0.02
0.04
0.05
0.27
0.45
6.09
7.68

0.00
0.01
0.01
0.02
0.07
0.00
0.00
0.00
0.01
0.04
1.10
1725"
0.46
0.56
0.69
5.31
0.13
0.20
0.24
1.92
2.37
.0..05
12.26

0.02
0.03
0.03
0.08
1.05
0.01
0.01
0.02
0.09
0.28
0.01
1757
7
0.99
1.09
1.73
3.26
4.30
0.29
0.31
0.52
3.01
0.19
0.01
15.69

0.06
0.07
0.10
0.36
0.82
0.01
0.02
0.03
0.18
0.02
,0.00
1755"
8
1.26
1.94
7.10
24.73
0.67
0.57
1.14
0.76
1.23
0.04
0.0,0
3^741

0.07
0.11
0.41
2.98
0.12
0.03
0.05
0.05
0.08
0.00
D.,00
'375TJ
9
0.10
1.93
0.30
0.01
0.00
16.56
0.02
5.42
0.04
0.00
0..00
2O9

0.01
0.11
0.02
0.00
0.00
0.79
0.00
0.35
0.00
0.00
0.00
rrz9~

-------
                                             10-15
Table 10-2.(continued).
Secondary 804
                                   Receptor Regions
Source. .Regions
1
2
3
4
5
6
7
8
9
10
11
Total
Note: Totals
.Source, .Regions
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
A.
0.
1
01
02
00
00
00
00
00
05
09
00
.0.0
16
may not
2
0.02
0.05
0.02
0.00
0.00
0.01
0.00
0.09
0.15
0.03
0..0.0
0.35
agree
3
0.06
0.13
0.04
0.01
0.00
0.02
0.00
0.18
0.37
0.06
0.00
0.87
due to
0.
0.
0.
0.
0.
0.
0.
0.
0.
'0.
0.
1.
4
05
10
07
05
03
03
02
08
23
49
.01 ,
16
5
0.03
0.05
0.04
0.07
0.08
0.01
0.02
0.03
0.08
0.11
0.36
0.88
6
0.10
0.19
0.19
0.19
0.39
0.07
0.08
0.11
0.32
0.22
.0.01
1.88
7
0.21
0.36
0.40
0.38
0.26
0.12
0.09
0.21
0.41
0.04
.0.00
2.48
8
0.24
0.56
0.94
1.46
0.05
0.20
0.17
0.29
0.22
0.01
.Q.J30
9
0.04
0.42
0.10
o.oo
0.00
1.10
0.01
0.86
0.02
0.00
.0..00
1755
rounding.
Receptor.
.Regions



  1 - Michigan

  2 - Illinois-Indiana

  3 - Ohio

  4 - Pennsylvania

  5 - New York to Maine

  6 - Kentucky - Tennessee

  7 - West Virginia to North Carolina

  8 - Rest of Eastern U.S. (Florida
      to Missouri to Minnesota)

  9 - Ontario

 10 - Quebec

 11 - Atlantic Provinces
1 - Boundary waters

2 - Algoma

3 - Muskoka

4 - Quebec

5 - Southern Nova Scotia

6 - Vermont-New Hampshire

7 - Adirondacks

8 - Pennsylvania


9 - Smokies

-------
                                           10-16

Table 10-3.  Transfer Matrix for July 1978 Sulfur Dioxide and Sulfate
             (Primary and Secondary) Concentrations  (ug/m3) from the
             ASTRAP Model Using the Phase II State/Provinces SC>2 Qnission
             Inventory and Primary Sulfate Emission Factors in Table 10-1.
S02
Source Regions
1
2
3
4
5
6
7
8
9
10
11
Total
Primary 504
Source Regions
1
2
3
4
5
6
7
8
9
10
11
Total
1
0.00
0.02
0.00
0.00
0.00
0.00
0.00
0.06
0.37
0.00
0.00
7756"

0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.01
0.02
0.00
0.00
"oTDT
2
0.05
0.15
0.02
0.00
0.00
0.00
0.00
0.18
1.03
0.00
0.00
"Ot

0.00
0.01
0.00
0.00
0.00
0.00
0.00
0.01
0.04
0.00
0.00
77DT
3
0.63
0.54
0.10
0.01
0.00
0.02
0.01
0.47
1.10
0.00
0.00
"278T

0.04
0.03
0.01
0.00
0.00
0.00
0.00
0.03
0.04
0.00
0.00
"oTTs
Receptor Regions
4-5 6
0.24 0.01 0.12
0.17
0.19
0.33
0.23
0.02
0.04
0.09
1.58
4.63
0.00
7738"

0.01
0.01
0.02
0.05
0.05
0.00
0.00
0.01
0.07
0.49
0.00
"Oo
0.02
0.06
0.21
0.39
0.01
0.05
0.01
0.16
0.33
3.40
1765

0.00
0.00
0.01
0.03
0.08
0.00
0.01
0.00
0.01
0.04
0.57
"0775
0.19
0.44
1.23
4.92
0.08
0.20
0.08
1.09
1.10
0.00
?75I

0.01
0.02
0.03
0.17
0.91
0.01
0.02
0.01
0.05
0.12
0.00
"OI
7
0.32
0.45
1.47
4.22
2.24
0.22
0.49
0.17
1.01
0.02
0.0,0
17760

0.02
0.03
0.09
0.42
0.41
0.01
0.03
0.12
0.06
0.00
0.00
T70S"
8
0.30
0.55
3.62
19.49
0.14
0.45
1.67
0.23
0.31
"o.oo
0.00
25776"

0.02
0.03
0.20
2.06
0.03
0.03
0.08
0.02
0.02
0.00
0..00
"OS"
9
0.00
0.07
0.03
0.00
0.00
12.52
0.01
3.39
0.00
0.00
0.0.0
lOT

0.00
0.00
0.00
0.00
0.00
0.54
0.00
0.22
0.00
0.00
0..00
7777

-------
Table 10-3.(continued).
                                          10-17
Secondary 804
1
2
3
4
5
6
7
8
9
'10
11
Total
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
ft
00
02
00
00
00
00
00
06
18
00
00
27
' 0
0
0
0
0
0
0
0
0
0
0
ft
.04
.12
.02
.01
.00
.00
.01
.16
.19
.00
.00
75T
0.19
0.35
0.09
0.01
0.00
0.03
0.01
0.33
0.18
0.00
0..00
tn$
*. '
0.12
0.21
0.28
0.29
0.12
0.03
0.07
0.12
0.76
0.69
0.00
"2768"
0.02
0.03
0.11
0.24
0.24
0.01
0.11
0.02
0.18
0.22
0.39
trsfr
0.11
0.23
0.47
0.73
0.85
0.13
0.25
0.11'
0.57
0.21
0.00
TT£6
0.19
0.38
0.95
1.18
0.25
0.28
0.40
0.21
0.36
0.01
0.00
~%rn
0.16
0.40
1.45
2.34
0.03
0.45
0.79
0.27
0.11
0.00
0.00
"Oo"
0.00
0.04
0.03
0.00
0.00
1.72
0.01
0.94
0.00
0.00
,0.00
•2T7T
 Note:  Totals may not agree due to rounding.

Source
  1 - Michigan

  2 - Illinois-Indiana

  3 - Ohio

  4 - Pennsylvania

  5 - New York to Maine

  6 - Kentucky - Tennessee

  7 - West Virginia to North Carolina

  8 - Rest of Eastern U.S. (Florida
      to Missouri to Minnesota)

  9 - Ontario

 10 - Quebec

 11 - Atlantic Provinces
     r. -Regions

1 - Boundary waters

2 - Algoma

3 - Muskoka

4 - Quebec

5 - Southern Nova Scotia

6 - Vernont-New Hampshire

7 - Adirondacks

8 - Pennsylvania


9 - Smokies

-------
                         CHAPTER 11




   CONCLUSIONS, RECOMMENDATIONS, AND PHASE III WORK PLANS






11.1 CONCLUSIONS AND RECOMMENDATIONS






     The review of sulfur chemistry concluded that the rate of



homogeneous gas phase conversion of S02 to 304 is dominated by



free radical reaction processes and the concentration of the



important free radicals are dependent on many factors, espec-



ially the concentration of volatile organic compounds and



nitrogen oxides, the temperature and the solar radiation in-



tensity.  Our knowledge of heterogeneous oxidation of S02 is



less complete, but indicates that liquid phase catalyzed oxida-



tion by manganese ion, carbon, and hydrogen peroxide all could



be potentially important.  However, there is uncertainty about



the actual availability of these catalyzing substances in ambient



fine particulate matter.  The review of nitrogen chemistry con-




cluded that the fate of nitric acid in the atmosphere is not well



understood, but it would still be useful to apply results of our



limited understanding of nitrogen chemistry to exploratory model-



ing exercises.  It was concluded from these modeling exercises



that greater temporal and spatial resolutions for nitrogen model-



ing is necessary and more extensive and reliable monitoring in-



formation is required for validation purposes.



     The review of the evidence for trends in precipitation com-



position and deposition concludes that, in spite of the difficul-

-------
                           11-2



ties with the data base and the controversy on the subject, the



the data do suggest expansion of the region covered by acidic



rainfall/ especially into the southeastern and mid-western



portions of the U.S. and the southeastern portions of Canada.



In this regard/ the Modeling Subgroup decided to focus its Phase



II efforts on the present situation and not to apply the models



to past data because of the uncertainties in past measurements



and other necessary input data, particularly the historic emiss-



ions fields.



     The review of the seasonal variation of deposition and



chemical transformation rates concluded that many of the para-



meters in regional models could be strongly dependent upon



latitude during the winter months and recommended that not only'



seasonal variability, but also the spatial variability be taken



into account.  This review provided some specific suggestions



for the participating models for Phase III.



     The review of the global distribution of pH and its impli-



cations concluded that all precipitation appears to contain both



acidic and basic material in small quantities.  There is evidence



that the "hemispheric background" value of precipitation pH (i.e.,



the average annual value at remote sites farthest removed from



the three major anthropogenic source regions in the northern



hemisphere) is significantly lower than the idealized "clean



atmosphere" value of 5.6.  This average hemispheric precipitation



pH value may now be closer to 5.0 than to 5.6.  More analysis



of data are required to determine what proportion of observed

-------
                             11-3
"background pH levels" are due to natural sources or to the



residual effect of man-made sources far upwind.



     The Phase II Data Analysis review found the highest preci-



pitation acidity on an annual basis in the northern hemisphere



over (1) eastern North America, (2) western Europe, and (3)



Japan.   Near neutral precipitation, frequently in excess of



pH 6, is found over the large continental areas of western



North America and Asia.  The cause of slightly acidic precip-



itation along the west coast of North America has not yet been



completely explained, but may be due to either anthropogenic



sources or the release of biologically-produced organic sulfur



compounds from the Pacific Ocean surface, or both.



     The zone of maximum acidity in eastern North America stretches



in a corridor through Ohio and Pennsylvania into Southern Ontario.



Available concentration data at remote locations in eastern North



America well removed from major source regions generally indicate



a summer sulfate maximum and a winter S02 maximum with highly



episodic behavior of both on a daily basis.  In addition, cal-



culated dry depositions of sulfur are found to be of comparable



magnitude to wet sulfur depositions especially close to source



regions and in the winter season.



     Recent interpretations of both concentration and deposition



monitoring data using trajectory calculations indicate that



maritime tropical air masses from the U.S. are the principal



conveyors of elevated concentrations and depositions to the

-------
                          11-4  '





extreme northeastern U.S. and southeastern Canada, as opposed to



continental polar air masses from Canada.  It was recoginized in



making these interpretations that source-receptor relationships,



based upon calculations of transport and chemical transformations



between probable sources and the receptor of interest and upon



event data at single monitoring stations, are not always straight-



forward and are subject to uncertainty.



     Finally, the data analysis review concluded that within east-



ern North America, natural sources of sulfur within the region



are unimportant compared to anthropogenic sources.  Somewhat more



significant are the background sulfur concentrations that are trans-



ported to eastern North America from the Pacific and Carribean



Oceans, the Atlantic Ocean south of 30°N, and the arctic region.



These manifestations of the hemispheric background contribution



to acid deposition are still considered to be small in comparison



with the local and long-range transport impacts in eastern North



America.  The predominant sources of elevated sulfur concentrations



in arctic air masses in the winter are thought to be, in order of



decreasing importance, Siberia, Europe, and Eastern North America.



     The Phase II analysis of the role of modeling indicated that



model predictions are expected to deviate to some degree from



actual monitoring measurements.  For practical reasons, models



do not incorporate all of our understanding of the relevant



physical processes, which itself is incomplete.  Furthermore,



our available monitoring data bases are insufficient to compute



the ensemble average which the model is designed to predict. The

-------
                           11-5
uncertainties in model predictions may be quantified from the



differences between model predictions and observations.



     Although the application of regional models is constrained



by these uncertainties inherent in their calculations, such



constraints can be alleviated to a significant degree by requir-



ing the modelers to quantify the relevant uncertainties, and by



taking these into account in any application.  Some of the un-



certainties in the transfer matrix elements can be assessed by



analyzing the transfer matricies from more than one model and by



using probablistic techniques of analysis which will be developed



during Phase III.  Other uncertainties may be quantified after



further model evaluation efforts are completed.



     While there is still no general agreement in the modeling



community as to (1) the proper method and (2) the statistics for



intercomparison and evaluation of models, the Modeling Subgroup



made significant strides in selecting a common basis for perform-



ing these tasks for the eight participating models.  In Phase II,



a complete set of evaluation statistics were computed by only



one modeler, while the monthly and annual residuals were computed



at 9 to 20 sites by three of the eight modelers.  The other four



modelers are expected to complete this minimum evaluation by



September 1981.  No single model has emerged as cleary superior



or inferior to the others from this first of three rounds of



model evaluation.  The evaluation has primarily served to reveal



(1) the deficiences in the monitoring data bases, (2) the need

-------
                           11-6 '





for some changes in input parameters for some of the models,



and (3) the need to use at least one more year of independent



data for model evaluation.



     The seasonal transfer matrices computed with several of



the models were too preliminary to draw any general conclusions/



but indicate that seasonal variability can be significant and



should be investigated further during Phase III.  In addition,



most of the models completed detailed sensitivity analyses by



varying each input parameter separately within the range normally



used for long-range transport modeling with either (1) actual



meteorology and emissions or (2) a hypothetical source-receptor



situation and simulated meteorology.  These sensitivity analyses,



which are documented in the individual Model Profiles, provide a



more complete understanding of the workings of'each model and



are useful for incorporating uncertainty in the analysis process.



     The annual transfer matrices for Phases I and II were inter-



compared and, interestingly, the use of standardized inputs did



not reduce the range of variation among models in some of the



transfer matrix elements expressed in two of the three standard



forms, namely, absolute values or normalized by unit emissions



of sulfur.  However, the transfer matrix elements expressed as



percentage contribution from a source region to a receptor area



were generally in much better agreement among the models.  Since



additional refinements will be made to most of the models and



a new set of source-receptor regions will be used during Phase



III, it is premature to draw any general conclusions at this

-------
                           11-7 .

time.  These variations in transfer matrix coefficients reflect
the current uncertainties in how best to parameterize all the
physical processes and are the result of different approaches by
independent modelers.  Although the desirability of a single/
unified transfer matrix was recognized, the Modeling Subgroup has
reservations about the generation and application of a unified
transfer matrix at this time because no matter how it is generat-
ed its interpretation is subject to some question.
     Since the modeling of nitrogen oxides chemistry is in its
infancy and because there is so little data available from which
to select the model parameterizations, any modeling at this stage
cannot yield more than some educated "first estimates" of the
long-range transport of nitrogen oxides. The preliminary results
from three models, in terms of transfer matrices and comparisons
to monitoring data, indicate the primitive nature of current
efforts.  Therefore, at this stage, these should not be used
for analysis efforts.
     A preliminary transfer matrix for primary sulfate emissions
was generated during Phase II to assess the relative contribu-
tion to acid deposition of primary sulfate emissions from oil-
fired and coal-fired combustion sources compared to that from
secondarily formed sulfates.  Newly published primary sulfate
emission factors from large package boilers were utilized by
one regional transport modeler to develop this comparative
analysis, the analysis indicated that primary sulfate concen-
trations exceeded secondary sulfate concentrations in the winter

-------
                           11-8
season, while secondary sulfate concentrations exceeded primary



sulfate concentrations in all other areas in the winter and in



all areas during the summer. Additional modeling of the contribut-



ion of primary sulfate emissions will be performed during Phase III



using an improved emissions inventory and an attempt will be



made to evaluate the results against monitoring data.

-------
                              .11-9
11.2  PHASE III WORK PLAN


       The principal objective for Phase III is to refine and

consolidate the information provided to the Coordinating

Committee at the end of Phase II.  Some attention will be

directed to additonal transboundary air pollution issues that

are likely to be considered in the Air Quality Agreeemnt

negotiations.

     The integrated Phase III report should then provide the

Coordinating Committee with substantially all the available

technical information and analysis relevant to closing nego-

tiations on a bilateral, transboundary air pollution agreement.

     Before proceeding to a consideration of specific tasks

for Phase III, it is useful to review briefly the direction

provided by the Coordinating Committee and the Work Group 2

progress to the end of Phase II.  The Coordinating Committee

recommended that Work Group 2 consider undertaking the following

tasks as early as possible in Phase II:

     o  Provide a means to estimate short-range and mesoscale
        transport for sulfur compounds relative to long-
        range transport for identified sensitive areas.
        Provide a means for evaluating such transport, if
        significant; and

     o  Assess the relative contribution to acid deposition
        on identified sensitive areas of primary sulfate
        emissions from oil-fired and coal-fired combustion
        sources in comparison with secondary formed sulfate
        from these sources.  Compare the primary sulfate
        deposition in identified sensitive areas from oil-
        fired sources with the total sulfur depostion from
        all other sources.

-------
                             11-10


     The Coordinating Committee also requested that in addition

to completing the acid deposition analyses initiated in Phase

I, Phase III work programs analyze other important trans-

boundary air pollution issues.  Among those issues of a

regional nature which are recognized to have an important

transboundary component are:

     o  Regional scale formation and transport of photochemical
        oxidants

     o  Deposition and concentration of heavy metals and organics

     In Phase II, Work Group 2 completed the following in

response to the above requests:

     1. provided a list of mesoscale models;

     2. estimated what proportion of the problem was due to
        the short, the medium and the long-range transport
        scales;

     3. provided an initial evaluation of primary sulfate
        emissions and their impact; and

     4. provided some insight into the assessment iteration*
        process

     In Phase III (June 1981-January 1982), Work Group 2 will

do the following:

     1. continue its evaluation and application of long-term
        regional models and consolidate the results;

     2. enlarge the modeling domain to include the entire
        continental U.S. and western Canadian provinces;
* The Assessment iteration process involves repeatedly analyzing
  transfer matrices, target depositions, and emissions and cost
  vectors as they relate to possible control strategies.

-------
                             11-11
     3. give additional emphasis to the atmospheric science
        review and monitoring data analysis efforts;

     4. run the selected (and other if appropriate) models
        using the unified S02 and NOX emission inventories
        on a state (multi-state)/province (sub-province)
        basis and unify, if possible, the transfer matrices;

     5. further quantify the "background S02 and NOX contribu-
        tions" to the observed concentrations and depositions;

     6. run the selected models on as many additional periods
        of meteorological data as possible to assess the vari-
        ability of the transfer coefficients (on a seasonal
        and annual basis)insofar as possible;

     7. propose a detailed work plan for the period beyond
        Phase 3 on additional transboundary air pollution
        issues as required by the Coordinating Committee;

     8. address the issues raised by the peer reviews and
        the assessment iteration process; and

     To carry out its tasks in Phase III, the Working Group

has established the sub-group and activity coordination scheme

illustrated in Figure 11-1.  Each sub-group will hold workshops

during Phase III.

     An updated work schedule for Phase III is presented in

Figure 11-2.

-------
                        11-12
Figure 11-1.  Phase III Organization for Work Group 2,
                      Co-Chairs

                      H. Ferguson
                      L. Machta
                    Vice Co-Chairs

                   G. Van Volkenburg
                   L. Smith
                     Co-Technical
                     Coordinators

                     J. Young
                     B. Niemann
1
Regional
Modeling
Sub-Group
Co-Coord inators
K. Demerjian
P. K. Misra








Monitoring and
Interpretation
Sub-Group
Co-Coord i nators
J. Miller
L. Barrie








Atmospheric
Science
Review
Sub-Group
Co-Coordinators
P. Altshuller
P. Sunimsrs








Local Source
Analysis
Co-Coord inators


To be selected
R. Shaw


-------
Figure 11-2.   Proposed Work Group 2 Phase III Activity Schedule  (Revised  May 15,  1981).
ACTIVITY
1. REGIONAL MODELING SUB-GROUP

Id. Ifeciinlcdl pctii. tjL"Gup f&vitiW— — — — — — — — — — — — — —
JLU. MJilSO-LlQclCG tJ:anSJ.ejr JUauiXX WOiTK ^annUaJ./ SeaSOnaJ. ) — — »

la • OtnGir pollutant nioctelincj 33 nGCGSSQiry cind f GcisiblG ~~ —
l€ • Rsspona to PGGIT JTGVIGWS ~-~~ ~* — _ — . — .
2. ATMOSPHERIC SCIENCE REVIEW SUB-GROUP

metals, etc.

2d. Review any new information on issues connected with Sulfur 	
and Nitrogen
2f. Analysis of initial comments on modeling, trends etc 	
9n Wfii-p enhni'Vrtin V£*rv»r-t-— — 	 — 	 — 	 — 	 — — 	 — 	 — 	 — — 	 __J-_ _— __ 	
July 81

x

"~~™ X






X


	 X

A
	
Aug. 81





—x



— — X



	
Sept 81
x

X
— x


X
X


X

	
Oct. 81
X





— x

Nov 81







Dec 81
x


X
X
X

Jan82
x
..
X
X

                                                                                                                            I
                                                                                                                            U)

-------
Figure 11-^^ontinued).    Proposed Work Group 2 Phase III
Schedule (Revised May 15, 1981).
ACTIVITY
3. MONITORING INTERPRETATION SUBGROUP

3a. Continental background— __— _ — — — —

Jc. kJGdsonaJL variations™ —- — — — —.——.—.———.——— — _«._.
JQ. L.iassi£ ication oy air mass origin etna episode cinuiyscis —
3e. pH maps analysis ~ — ~ ——

3g. Analyze repot ts of anamolous or confounding roonitoring data— ~
4. IJOCAL SOURCE ANALYSIS SUB-GROUP

deposition
to transboundary effects

analysis of data
*ie« wtiuc oui,>*yjroup j-trpoi. T. ~ — -— — — - .— — — _«._«. ._- M — — _ «._._ —
5. W3RKING GROUP

Dti» Lxrcrput uuion or J.nucg.LUutX4 L ±nuj. i(jpoi.c — -...,-.. .— .- ....-...._-_— _ — ...»,.«
July 81

X



X














Aug. 81













—x




SeptS 1
X
X
X


X
X


X
X
X



	 X

Oct. 81






*

Nov. 81



X



Dec. 81
X
X

X



Jan82
»-
H-
H
.£
X
X
X


-------
                           Chapter 12

     PRELiM.i.NAJR.Y,' _PfiQ.P.6.SA.LS '.FQR' RESEARCH., ' MQD.E.L.I.N.G., .AND.'
            M.Q.NJ. TORI ' NG. ELEMENT. .OF. Tfl.E.
12.1  ,1 At ic od .u.c.t ,ip n
     A transboundary air pollution agreement between the
United States and Canada that  reflects our scientific under-
standing in the early  1980s  of this problem may not be sufficient
to resolve completely  all aspects  of the problem.  Therefore,
it is important to provide a mechanism whereby new knowledge
can be acquired and assimilated by those responsible for
managing the environmental protection programs of the two
countries.  Additionally, recognizing our limitations to
predict with certainty the full consequences of controlling
emissions, it is important to  determine by observations of
concentrations and deposition  rates the effectiveness of the
initial agreement in resolving transboundary impact issues.
     The following four subsections suggest areas where uni-
lateral and cooperative programs aimed at providing increased
understanding could further  this objective.  They have been
categorized under two  headings. First, E,s,s,en.t,ial items are
those that are considered to be necessary for improved
understanding in the near term (next .few years).  It is
recommended that the programs  of both governments include
sufficient resources to adequately fund these activities.
Second, .Desirable items are  those  that are considered useful

-------
                             12-2





for improved understanding in the short-term/ or may be



essential to the long-term resolution of some transboundary



issues.  In making the following recommendations the Work



Group has taken into consideration current U.S. and Canadian



planning information as embodied in the U.S. National Acid



Precipitation Assessment Plan and the Canadian LRTAP plan.



The Work Group recommends that a bilateral air quality res-



earch and monitoring committee be established to coordinate



these activities beyond January, 1982.



12.2  Atmospheric Processes Research



Essential



     0    scavenging mechanisms for wet and dry removal of



          transported pollutants.



     0    background concentrations and ventilation losses



          at ground level and aloft.



     0    atmospheric chemistry for acid formation and



          neutralization processes, especially those for



          nitrogen chemistry



     0    improved understanding of long-distance meteorological



          transport.



     0    prediction of meteorological conditions associated



          with transboundary air pollution.



Desirable



     0    reconstruction of transport and dispersion during



          episodes

-------
                             12-3
     0    global background of acids and precursors — concentra-
          tion/ extent, frequency and their explanation.
     0    cooperative tracer studies.
     0    atmospheric chemistry of oxidant precursors and their
          aged reaction products.
     0    vertical sulfur concentrations and regional horizontal
          flux studies.
12.3  Monitoring
Essential
     0    development of improved acid deposition monitoring
          devices, particularly for dry deposition.
     0    development of uniform protocol for acid deposition
          monitoring procedures, sample handling and analysis,
          data archiving, quality assurance, and exchange of
          data sets between the US and Canada.
     0    coordination of gaseous and particulate pollutant
          monitoring for establishing air quality trends.
     0    sustained monitoring program for acid deposition and
          air pollutant concentrations in North America for:
             remote sites potentially affected  and sensitive
             areas currently impacted.
             along the Canada/U.S. border.
     0    monitoring fluxes crossing the periphery of the
          modeling region.

-------
                             12-4

Desirable
     0   . feasibility study on the establishment of cooperative
          episode warning network for harmful pollutants.
     0    assessment of needs for, and as appropriate/ imple-
          mentation of,
          -  a heavy metals monitoring network.
          -  an oxidants monitoring network.
             a toxic organics monitoring network.
             a supplementary visibility monitoring network.
12.4  Modeling
Essential
     0    study approaches for providing improved quantification
          of the uncertainties associated with modeling source-
          receptor relationships.
     0    continue validation and application of available
          models to the North American situation so as to
          provide increased understanding of source-receptor
          relationships.
     0    development of a comprehensive North American acid
          deposition model.
Desirable
     0    development of a regional oxidant model for application
          in eastern North America.
     0    development of regional haze model for application
          in eastern North America.

-------
                             12-5





     0    development of a heavy metals atmospheric transport



          model.



     0    development of a toxic organics atmospheric transport



          model.



     0    development of techniques for predicting the occurrence



          in real time of transboundary air pollution episodes.



12.5  Atmospheric Science Assessment



Essential



     0    produce a regular(annual) bilateral report which updates



          our current understanding of transboundary air pollution



          phenomena, including our ability to monitor it and model



          its source-receptor relationships.  (A continuing bi-



          lateral committee could be charged with this task.)



     0    produce critical reviews/ as warranted, of major



          developments affecting our understanding of trans-



          boundary air pollution.

-------
                               R-l


                            REFERENCES
Barrie, L. A., et al., 1980:  The Canadian Air and Precipitation
  Monitoring Network APN, in Atmospheric Pollution 1980, Proceedings
  of the 14th International Colloquium, Paris, France, May 5-8,
  Studies in Environmental Science, Volume 8, Elsevier, Amsterdam.

Barrie, L. A., 1981:  Environment Canada's Long Range Transport of
  Atmospheric Pollutants Program:  Atmospheric Studies, Paper
  Presented at the Conference on the Effects of Acid Precipitation
  On Ecological Systems in the Region of the Great Lakes, March 31-
  April 3, Michigan State University.

Barrie, L. A., 1980:  The Fate of Particulate Emissions from An
  Isolated Power Plant in the Oil Sands Area of Western Canada,
  Annals. New York Acad. Sci., 338, 434-452.

Barrie, L. A., et al., 1981:  The Influence of Mid-Latitudinal
  Pollution Sources on Haze in the Canadian Arctic, Atmos.
  Environ., 15, in press.

Barrie, L. A., et al., 1981:  The Temporal and Geographical
  Distribution of Acidic Pollutants in the Atmosphere of Eastern
  Canada, Proc. Conf. on Long-Range Transport of Airborne Pollutants,
  April 29-May 1, Albany, New York.

Barrie, L. A., 1981:  The Prediction of Rain Acidity and S02
  Scavenging in Eastern North America, Atmos. Environ, 15,  31.

Beilke, S., 1970:  Laboratory Investigations of Washout of
  Trace Gases, AEC Symp. Ser.  Proc. Symp. on Precip.
  Scavenging.

Benkeley, C. W. and Mills, M. T. 1980:  Development and Validation
  of a Simple Regional Climatological Dispersion Model for Nitrogen
  Oxides, Report Prepared for Argonne National Laboratories by
  Teknekron Research, Inc.,  November.

Benkowitz, C., 1979:  Compiling a Multistate Emissions Inventory,
  Brookhaven National Laboratory/MAP3S Report, BNL-26843,
  Brookhaven National Laboratory, Upton, N.Y.

-------
                               R-2
Bolin, B. and Persson, C., 1975:  Regional Dispersion and
  Deposition of Atmospheric Pollutants with Particular Applica-
  tion to Sulfur Pollution over Western Europe, Tellus, 27,
  281-310.
Bottger, et al., 1978.  Atmospharische Kreislaufe van
  Strichoxiden and Ammoniak, Kernforschungsaulage Julich GmbH,
  Inst. Chemie:  Jul-1558.

Botthheim, J. W. , 1981:  Discussion on the Incorporation of
  Oxides of Nitrogen Chemistry in LTRAP Models, Atmoshperic
  Environment Service Internal Report.

Chung, Y. S., 1977:  Sources and Sinks of Photochemical Ozone
  (03) in the Boundary Layer, Proceedings of the Fourth
  International Clean Air Congress, Tokoyo, 138.

Clark, T. L., 1980:  Annual Anthropogenic Pollutant Emissions in
  the United States and Southern Canada East of the Rocky
  Mountains, Atmos. Environ., 14, 961-970.

Reisinger, L.M. and Crawford, T. L., 1981:  Interregional
  Transport:  Case Studies of Measurements Versus Model Predictions,
  Paper 81-13.4 presented at the 74th Annual Meeting of the APCA,
  June 21-26, Philadelphia, PA.

Fay, J. A. and Rosenzweig, J. J. 1980:  An Analytical Diffusion
  Model for Long Distance Transport of Air Pollutants, Atmos.
  Environ., 14, 355-365.

Galloway, J. N., and Whelpdale, D. M, 1980:  An Atmospheric
  Sulfur Budget for Eastern North American, Atmos. Environ., 14,
  409-417.

Gravenhorst, G., et al., 1980:  Sulfur Dioxide Absorbed in Rain
  Water, in Proceedings of the Effects of Acid Precipitation on
  Terrestrial Ecosystms, Plenum Press, New York.

Hales, J. M. 1981:  OSCAR and Miscellany, Directions Memo, MAP3S/
  RAINE Program, June 1, 6 pp.

Henry, R.C. and Hidy, G. M. , 1980:  Potential for Atmospheric
  Sulfur from Microbiological Sulfate Reduction, Atmos Environ.,
  14, 1095-1104.

-------
                                 R-3

Hicks, B.B.  and  Shannon, J. D. , 1979:  A Method  for Modeling the
  Deposition of  Sulfur by Precipitation Over  Regional Scales,
  .J,./ Appl,.; M.e.teror..' ,  X8, 1415-1420.

Homolya, J.  B. and Lambert, S. 1981:  A Charactrization of
  Sulfate  Emissions  from Non-Utility Boilers  in  New York City Firing
  Low-Sulfur Residual Oils, .Ji_\Air."_jPQ,ll.|,"Qqp,t,j_.'_Aggp.c•_ i  31,
  139-143.

Huschke, R.  E.,  1969:  Arctic Cloud Statistics  from 'Air
  Calibrated1  Surface Weather Observations, U.S.  Air Force
  Project, Rand  Contract No. F 44620-67-C-0015.

Jeffries,  H.  E.,  1979:  Use of Laboratory  Data  to Describe
  the Fate of  NOX in  the Atmosphere, Presented  at the Technical
  Symposium  on the Implications of a Low NOX  Vehicle Emission
  Standard,  Reston,  Virginia, May 2-4, 1979.

Johnson, W.  B.,  et al., 1978.  Long-Term Regional Patterns and
  Transfrentier  Exchanges of Airborne Sulfur  Pollution  in Europe,
  .A.tmpS,..'.E.nyi.rqn.yf .1.2, 511-527.

Koerner, R,  M. and Fishe,D. 1981:  Acid Snow  in  the Canadian
  High Arctic, submitted to Nature,

Lamb, R. G.,  1980:  Mathemetical Principles of Turbulent Diffusion
  Modeling,  Article  in Atmospheric Planetary  Boundary Layer
  Physics, Elsevier  Publishing, Amsterdam.

Levine, S. Z.  and Schwartz, S. E, 1981.  In-cloud and Below-
  Cloud Scavenging of Nitric Acid Vapor, A.tmp,s.., Eny.i.rpn. ,
  15, in press.

Lyons, W.  A.,  1980:   Evidence of Transport of Hazy Air  Masses
  from Satellite  Imagery,' Annal,., N..Y... A.c.ad... S.cj... , .33,8,  418-433.

MAP3S/RAINE,  1981:  The MAP3S/RAINE Precipitation Chemistry
  Network:   Statistical Overview for the Period  1976-1980,
  submitted  to Atmp.s.y' .Eny.iy.on •

Merritt, W.  F, 1976:   Trace Element Content of Precipitation
  in a Remote  Area,  in Measurement.,, Pe.te.cti.oa^ a^nd, .Cpntrp.!
  p_f._,E.ny.i.r.o.nm.e.n.t.a.l,'_P.Q 11 At.an.ts, Internationa 1  AtomTc Energy
  Agency 7  Vienna7 7 5-87.

-------
                              R-4
Mueller, P. K., et al., 1979:  Some Early Results from  the
  Sulfate Regional Experiment  (SURE), in Proc. of the 4th
  Symposium on Turbulence, Diffusion and Air Pollution, January
  15-18, Reno, Nevada, American Meterological Society,  Boston, MA.

Niemann, B. L., et al., 1980:  Initial Evaluation of Regional
  Transport and Subregional Dispersion Models for Sulfur
  Dioxide and Fine Particulates, Proceedings of the Second
  Joint AMS/APCA Conference on Applications of Air Pollution
  Meteorology, March 24-27, New Orleans, LA.

Niemann, B. L., et al., 1979:  Application of a'Regional Trans-
  port Model to the Simulation of Multi-Scale Sulfate Episodes
  Over the Eastern United States and Canada, in Proceedings of
  the WMO Symposium on the Long-Range Transport of Pollutants
  and Its Relation to General Circulation Including Stratospheric/
  Troposheric Exchange Processes, October 1-5, Sofia, Bulgaria.

Olson, M. P., et al.  1979:  A Concentation/Deposition  Model
  Applied to the Canadian Long-Range Transport of Air Pollutants
  Project, LRTAP 79-5, Atmospheric Environment Service, Downsview,
  Ontario.

Paresh, P. P. and Husain L., 1981:  Windflow Patterns and
  Particulate Sulfate Concentration in Ambient Air at Whiteface
  Mountain, New York:  A Continuous 18 Month Inventory, submitted
  to Science.

Patterson, D. E., et al. 1981:  Monte Carlo Simulation  of Daily
  Regional Sulfur Distribution:  Comparison with SURE Sulfate
  Data and Visual Range Observations During August 1977, j,..'
  APPl... M.e.t.epf,* , ,20, 70-86
Peterssen, S. 1956:  .Weather, .Analysis, .ap.d. fpre.c.as.tipg, McGraw-
  Hill, New York.

Portelli, R. V. , 1977:  Mixing Heights, Wind Speeds and Ventilation
  Coefficients for Canada, Climatological Studies No. 31,
  Atmospheric Environment, Downsview, Canada.

Rutherford, I. D., 1977:  An Operational Three Dimensional
  Multivariate Statistical Objective Analysis Scheme, Issue  #1,
  Notes Scientifiques at Techniques, RPN, Atmospheric Environment
  Service, Dorval, PQ.

Samson, P. J., 1980:  Trajectory Analysis of Summertime Sulfate
  Concentrations in the Northeastern United States. ,»J..__.'_Appl..
  Meteor-, 1.9, 1382-1394.

-------
                               R-5
Shannon, J. D.,  1979:  The Advanced Statistical Trajectory
  Regional Air Pollution Model, in Proceedings of the Fourth
  Symposium on Turbulence, Diffusion, and Air Pollution, January
  15-18, Reno, Nevada, American Meteorological Society,
  Boston, MA.
                       •
Shannon, J.D., 1981:  A Model of Regional Long-Term Average
  Sulfur Atmospheric Pollution, Surface Removal, and Net
  Horizontal Flux, Atmos. Environ,  15, 689-701.

Shieh, C. M., et al., 1979:  Estimated Dry Deposition Velocities
  of Sulfur Over the Eastern United States and Surrounding Regions,
  Atmos. Environ., 13, 1361-1368.

Shenfeld, L. , et al., 1980:  An Air Pollution Incident in Southern
  Ontario, Canada with International Implications, Ontario
  Ministry of the Environment, Toronto, Ontario.

Slinn, W. G. N., et al., 1979:  Wet and Dry and Resuspension of
  AFCT/ TFCT Fuel Processing Radionuclides, U.S. Department of
  Energy Final Report SR-0980-10,  Air Resources Center, Oregon
  State University, Corvallis, Oregon.

Sykes, R. I. and Hatton, L. , 1976:  Computation of Horizontal
 ' Trajectories Based on the Surface Geotrpphic Wind, Atmos.
  Environ., 10,  925.

Szabo, M. F., et al., 1981:  Perspectives on the Issue of Acid
  Rain, Final Draft Report to the U.S. Department of Energy,
  Morgantown Energy Technology Center, Contract No. DE-AC21-
  81MC16361, June.

Tennekes, H., 1977:  The General Circulation of Two-Dimensional
  Turbulent Flow on a Beta plane,   J. Atmos. Sci., 34, 702-712.

Voldner, E. C.,  et al., 1980:  A Preliminary Canadian Emissions
  Inventory for Sulfur and Nitrogen Oxides, Atmos. Environ., 14,
  419-428.

Voldner, E. C.,  et al., 1980: Comparison Between Measured and
  Computed Concentrations of Sulfur Compounds in Eastern North
  America.  AQRB-80-0003-T (LRTAP-02), Atmospheric Environment
  Service, Downsview, Ontario (To be published in Journal of
  Geophysical Research, 1981).

-------
                              R-6
Walmsley, J. L., et al.,  1981:   Sensitivity  Tests  with  a Trajectory
  Model/ to be published  in the  Air  Quality  Research  Branch
  Report Series, Atmospheric Environment  Service,  Downsview.

Whelpdale, I. M. 1978:  *jaiT9e~?9a-'-e  Atmospheric  Sulfur  Studies
  in Canada, Atpips...'Ep.v;..rpp.., .1,2,  661-670.

Wilson, J. et al., 1980:  Wet Deposition  in  the  Northeastern
  United States, Atmospheric Sciences  Research Center Publication
  796, State University of New York  at Albany, December.

Wilson, W. E/ and Gillani, N. V.,  1979:   Transformation During
  Transport:  A State-of-the-Art Survey of the Conversion of  S02 to
  Sulfate, WMO Symposium  on the  Long-Range Transport  of Pollutants
  and Its Relation to General Circulation Including Stratospheric/
  Tropospheric Exchange Processes, 1-5 October,  Sofia,  Bulgaria.

-------
                          ADDENDUM A




                 Model Critique by L. Machta


               U.S. Co-chairman of Work Group 2
           •9

     The models in this report which provide the transfer



matrices represent the cutting edge of the state-of-the-art


in acid rain modelling.  They generally incorporate in a


practical way the best available physics, chemistry and



meteorology.   Their products have been tested/ insofar as


possible/ against many real atmosphere measurements.  Thus,


the criticism below is not directed at deliberate misconceptions



or imperfections of acid rain models, but rather at the current


inadequate state-of-the-art.


     It is this writer's judgment that the acid rain models


are not yet good enough to use their transfer matrices to


play a role in assessing emission controls to reduce acid


rain impacts.  The reasons for this inadequacy are noted below:



     1.  Current acid rain models are currently-capable of


predicting only the sulfate and SC>2 deposition and air concen-


trations.  They cannot, as yet, directly predict deposition


or concentration of hydrogen ions, NOX, NO-j", and other major


ions like ammonia, calcium or magnesium that bear on acid rain.


     2.  The models do not account for the low pH in isolated


regions of the globe where the precipitation often is almost


(or just) as acid as downwind of large anthropogenic sources.

-------
                            A-2


The presence of this global acid rain means either that there

are larger natural sources or longer distance transport from

man-made sources than implied by the present models.

     3.  The flux to the ground, vegetation, water, etc.

cannot now be adequately measured.  This leaves the budgets

of acid material in doubt and precludes complete model valida-

tion.  It also prevents a correlation between total acid

deposition and the observed impact in the field.

     While the above represent the writer's main reservations,

some critics have noted other factors which also possess

merit such as: the sulfate chemistry in the models which is

simplified may not adequately represent real atmospheric
                                                           •
chemistry; transport and dispersion calculations over long

distances have not been validated; and precipitation scavenging

is also simplified based on very limited data.

     To summarize, the criticism of models applies mainly to

the inadequate state of knowledge rather than the model's

incorporation of known information.  This writer does not

believe that they should be used as yet to assess the benefits •

of emission restrictions.  The above, however, should not be

interpreted as contending that no benefit will accrue from

fewer emissions of man-made sulfur oxides, but rather that the

models are not yet ready to quantitatively predict the reduc-

tion in acid deposition.

-------
                          ADDENDUM B


             Response to Dr. Machta's Addendum A

                      by H. L. Ferguson
             Canadian Co-Chairman of Work Group 2    <


     Dr. Machta's critique identifies some aspects of long

range transport modeling where virtually all modelers

recognize that our knowledge is incomplete.  Much of the

current and planned activity of Work Group 2 is aimed at

reducing the uncertainties he describes.

     The basic question of whether or not the transfer matrices

are good enough to "play a role in assessing emission controls

to reduce acid rain impacts" is a matter c?f "shades of grey"

judgment.  At what_point does such a synthesis of scientific

model estimates (with their implied error bands) become "good

enough"?  Canadian modelers generally hold the view that the

transfer matrices provide sufficiently indicative information

linking causes-and effects between larger areas (such as

groups of states)  to be used for implementing initial control

actions.  This view appears to be shared by many European and

U.S. modelers.

     Some general philosophical comments on the present stage

of LRTAP models and their potential applications are presented

in Chapter 4 of this report.

     With reference to the three specific points raised by

Dr. Machta, I would like to make the following observations:

-------
                        B-2

1.   The capability of estimating hydrogen ion concentra-
     ion is particulary important because of its direct
     relationship with pH.  Models can predict deposition
     and concentrations of hydrogen ions indirectly using
     various methods.  The sulfate surrogate method/ for
     example, reproduces very well the H+ concentration
     in Eastern North America (Barrie, 1981).
     Preliminary linear nitrogen and nitrate modeling
     is presented in Chapter 9 of this report and is
     discussed by several of the modelers.
     Modeling of other major ions has not been attempted
     to date because the required inventories do not
                   »
     exist.
2.   Models have not been designed for and have not been
     applied to the prediction of pH in isolated regions
     of the globe, so we cannot conclude anything about
     our skill in predicting in these areas.  We do know
     that in such areas the data base is grossly inade-
     quate for developing and testing LRT models.  We
     also know that the "low" average precipitation pH
     values being reported at a few remote locations still
     represent 5 to 10 times less hydrogen ion than average
     values in the most affected area of eastern North
     America.
     Very long range transport does occur and has been
     widely reported.  A few examples are:
     (1)  Large uncontrolled forest fires in Western

-------
                        B-3



          North America in the early 1950's caused a


          blue haze over Europe for many weeks;


     (2)  Recent measurements coupled with back trajec-


          tories in the Arctic have indicated sources in
              i

          the USSR, China and Eastern North America


          (Barrie,  1981) ; and


     (3)  Preliminary analysis of the Bermuda wet

          chemistry data indicate that low pH episodes


          have their origin in Eastern North America


          while higher pH episodes are associated with


          long sea  trajectories from the south and east

          (personal communication from John Miller).


3.*   All flux components (wet and dry) can and have been


     measured although the error bands are large in some


     cases.   Dry flux is not, however, measured routinely


     and is,  therefore, not available for validation


     purposes.

     Intercountry sulphur budgets have been independently


     determined by  various authors using various techniques,


     in both  the USA and Canada, and are generally in good


     agreement.


     The Working Group 1, Phase II data show an excellent

     correlation between measured excess sulfate in lake


     and river systems and modeled wet sulfate deposition

     from the models, indicating that only a small fraction

-------
                             B-4





          of the dry deposited sulphur becomes free hydrogen



          ion in an aquatic system.  This allows modeled



          results to be tied directly to effects in the



 t         aquatic ecosystem.



          Plans call for a high priority to be given to more



        •  integration of the Working Group 1 and 2 products



          in Phase III.



     We accept that there is uncertainty in all measured and



modeled parameters.  I think we can better define the limits



of confidence during Phase III.  There is no point, however,



in improving the chemistry, or any other parameterization, to



produce changes in output that are insignificant compared to



the errors in the measured data.



     I do not personally believe that the models are misre-



presenting the major features of the acid deposition problem



in eastern North America.



     A comparison of available hemispheric data on acid



depositions with major hemispheric emission regions of acid



precursors shows that sensitive pristine regions are being



adversely affected by source regions hundreds to thousands of



kilometers distant.



     Current models and transfer matrices show the general



source-receptor relationships involved.   They represent the



best method available for evaluating the potential ameliorative



effects of alternative control strategy scenarios.

-------
                             B-5






effects of alternative control .strategy scenarios.  In my




view they provide sufficient information to make a start now



on a broad agreement for reversing the deterioration of the



environment due to acid deposition in Eastern North America.



As the models and transfer matrices continue to be improved



through substantial longer term research efforts in the



United States/ Canada and elsewhere, the initial broad agree-




ment on controls can and should be periodically reviewed and



refined.

-------
       APPENDIX 1

      Work Group 2

   Terras of Reference
and Additional Guidance

-------
                            A.1-1


               Terms of Reference from the MOI

    The Group will provide information based on cooperative

atmospheric modeling activities leading to an understanding

of the transport of air pollutants between source regions and

sensitive areas, and prepare proposals for the "Research,

Modeling and Monitoring" element of an agreement.  As a first

priority the Group will by October 1, 1980 provide initial

guidance on suitable atmospheric transport models to be used

in preliminary assessment activities.

    In carrying out its work, the Group will:*

       identify source regions and applicable emission

       data bases;

       evaluate and select atmospheric transport models

       and data bases to be used;

       relate emissions from the source regions to

       loadings in each identified sensitive area;

    -  calculate emission reductions required from source

       regions to achieve proposed reductions in air

       pollutant concentration and deposition rates which

       would be necessary in order to protect sensitive

       areas;
*  proposed additional term of reference:
     11 - evaluate and employ available field measurements,
         monitoring data and other information;"

-------
                             B-4





          of the dry deposited sulphur becomes free hydrogen



          ion in an aquatic system.  This allows modeled



          results to be tied directly to effects in the



      +   aquatic ecosystem.



          Plans call for a high priority to be given to more



          integration of the Working Group 1 and 2 products



          in Phase III.



     We accept that there is uncertainty in all measured and



modeled parameters.  I think we can better define the limits



of confidence during Phase III.  There is no point, however,



in improving the chemistry, or any other parameterization, to



produce changes in output that are insignificant compared to



the errors in the measured data.



     I do not personally believe that the models are misre-



presenting the major features of the acid deposition problem



in eastern North America.



     A comparison of available hemispheric data on acid



depositions with major hemispheric emission regions of acid



precursors shows that sensitive pristine regions are being



adversely affected by source regions hundreds to thousands of



kilometers distant.



     Current models and transfer matrices show the general



source-receptor relationships involved.   They represent the



best method available for evaluating the potential ameliorative



effects of alternative control strategy  scenarios.

-------
                             B-5





effects of alternative control strategy scenarios.  In my



view they provide sufficient information to make a start now



on a broad agreement for reversing the deterioration of the



environment due to acid deposition in Eastern North America.



As the models and transfer matrices continue to be improved



through substantial longer term research efforts in the



United States, Canada and elsewhere, the initial broad agree-



ment on controls can and should be periodically reviewed and



refined.

-------
       APPENDIX 1

      Work Group 2

   Terras of Reference
and Additional Guidance

-------
                            A.1-1


               Terms of Reference from the MOI

    The Group will provide information based on cooperative

atmospheric modeling activities leading to an understanding

of the transport of air pollutants between source regions and

sensitive areas, and prepare proposals for the "Research,

Modeling and Monitoring" element of an agreement.  As a first

priority the Group will by October 1, 1980 provide initial

guidance on suitable atmospheric transport models to be used

in preliminary assessment activities.

    In carrying out its work, the Group will:*

       identify source regions and applicable emission

       data bases;^

       evaluate and select atmospheric transport models

       and data bases to be used;

    -  relate emissions from the source regions to

       loadings in each identified sensitive area;

       calculate emission reductions required from source

       regions to achieve proposed reductions in air

       pollutant concentration and deposition rates which

       would be necessary in order to protect sensitive

       areas;
*  proposed additional term of reference:
     " - evaluate and employ available field measurements,
         monitoring data and other information;"

-------
                        A.1-2






-  assess historic trends of emissions, ambient



   concentrations and atmospheric deposition to gain




   further insights into source-receptor relationships



   for air quality, including deposition; and



   prepare proposals for the "Research, Modeling and




   Monitoring"-element of an agreement.





    Additional Guidance from the Chairman of WG 3A



Each Work Group will be responsible individually for the



following:



a.  Develop data needs and analysis methods for their Work



    Group; identify required inputs from other Work Groups;




    (due to the size of the Work Groups, the Chairmen will



    have to very carefully orchestrate the Group's activities



    in order to accomplish their tasks).



b.  The technical review (including peer review as necessary)



    of their work products.



c.  Maintaining agreed upon work schedules with prompt



    notification to 3A Chairman in the event of any



    significant deviation from Work Plan.



d.  Responsible for coordination with their counterparts



    from the other country in conducting full cooperative



    analyses in order to fulfill the terms of reference.



e.  Responsible for fulfilling requests for information



    from other work groups in a timely fashion.

-------
                        A.1-3
f.  Be prepared to draft language for portion of agreement



    that pertains to their tasks as directed by Coordinating



    Committee.

-------
        APPENDIX 2




Membership of Work Group 2

-------
1.  United States
                            A.2-1
Chairman:
Vice Chairman;
Lester Machta, Director
Air Resources Laboratory, (Room 613)
National Oceanic and Atmospheric
  Administration
8060 13th Street
Silver Spring, MD  20910
(301) 427-7645

Lowell Smith, Director
Program Integration and
  Policy Staff (RD-681)
Environmental Protection Agency
Washington, D. C.  20460
(202) 426-9434
Members:
Paul Altshuller
Environmental Sciences Research
  Laboratory (MD-59)
Environmental Protection Agency
Research Triangle Park, NC  27711
(919) 629-2191

Franz Burmann
Environmental Monitoring Systems
  Laboratory (MD-75)
Environmental Protection Agency
Research Triangle Park, NC  27711
(919) 629-2106

Richard Harrington *
Morgantown Energy Technology Center
Department of Energy
Morgantown, West Virginia  26505
(304) 599-7529

Roger Morris
Office of Policy and Evaluation (PE-83)
Department of Energy
1000 Independence Avenue, S.W.
Washington, D. C.  20585
(202) 252-6453

-------
        A.2-2
Bernard  Silverman
Water  and Power  Resources  Services
E  & R  Center   P. 0. Box  25007
Department of  Interior
Bldg.  67 - Denver Federal  Center
Denver,  CO  80225
(303)  234-2576

Alternate for  Silverman

Richard  Ives
Department of  Interior,  Code 124
Washington, D.C.  20240
(202)  343-6703

John Ficke *
Council  of Environmental Quality
722 Jackson Place, N. W.
Washington, D. C.  20006
(202)  395-5760

Richard  Ball *
•Regional Impacts Division
Department of  Energy  (EV-24)
Washington, D. C.  20545
(301)  353-5801

Ken Demerjian
Meteorology and  Assessment Division  (MD-80)
Environmental  Protection Agency
Research Triange Park, NC  27711
(919)  629-3660

Nels Laulainen *
Office of Environmental  Processes
   and  Effects  Research (RD-682)
Environmental  Protection Agency
Washington, DC   20460
(202)  426-0803

Brand  Niemann
Program  Integration and  Policy Staff  (RD-681)
Environmental  Protection Agency
Washington, DC   20460
(202)  755-0324

Joe Tikvart
Office of Quality Planning and Standards  (MD-14)
Environmental  Protection Agency
Research Triangle Park,  NC  27711
(919)  629-5261

-------
                            A.2-3
                    Terry Clark *
                    Meteorology and Assessment Division (MD-80)
                    Environmental Protection Agency
                    Research Triangle Park, N.C.  27711
                    (919) 629-4524

                    John Miller
                    Air Resources Laboratory
                    National Oceanic and Atmospheric
                      Administration
                    8060 13th Street
                    Silver Spring, MD.  20910
                    (301) 427-7645

                    Jack Blanchard
                    OES/ENH   Room 7820
                    State Department
                    2101 C Street, N. W.
                    Washington, DC   20520
                    (202) 632-5748
Liaison:             Robin Porter
                    Department of State
                    EUR/CAN  Room 5227
                    2101 C Street, N.W.
                    Washington, DC  20520
                    (202) 632-3189
                    Dolores Gregory
                    Office of International Activities (A-106)
                    Environmental Protection Agency
                    Washington, DC  20460
                    (202) 755-0430
* Indicates new member of the Work Group

-------
2.  Canada
                            A.2-4
Chairman:
Vice Chairman;
 Howard Ferguson, Director
 Air Quality and Inter-environmental
   Research Branch
 Atmospheric Environment Service
 4905 Dufferin Street
 Downsview, Ontario  M3H5T4
 (416) 667-4937

 Greg Van Volkenburgh, Director
 Air Resources Branch
 Ontario Ministry of the Environment
 880 Bay Street, 4th Floor
 Toronto, Ontario, M5S1Z8
 (416) 965-6343
Members:
 Douglas M. Whelpdale
 Air Quality and Inter-environmental
   Research Branch
 Atmospheric Environment Service
   Environment Canada
 4905 Dufferin Street
 Downsview, Ontario, M3H5T4
 (416) 667-4785

James W.S. Young
Air Quality and Inter-environmental
  Research Branch
Atmospheric Environment Service
  Environment Canada
4905 Dufferin Street
Downsview, Ontario, M3H5T4
(416) 667-4786

Marvin P. Olson
Air Quality and Inter-environmental
  Research Branch
Atmospheric Environment Service
  Environment Canada
4905 Dufferin Street
Downsview, Ontario  M3H5T4
(416) 667-4903

Peter W. Summers
Air Quality and Inter-environmental
  Research Branch
Atmospheric Environment Service
  Environment Canada
4905 Dufferin Street
Downsview, Ontario, M3H5T4
(416) 667-4785

-------
                            A.2-5
                    Frank Vena *
                    Pollution Data Analysis Division
                    Environmental Protection Service
                      Environment Canada
                    Place Vincent Massey
                    Ottawa, Ontario, K1A1C8
                    (819) 997-3354

                    B . Power
                    Environmental Management
                      and Control Division
                    Newfoundland Department of Provincial
                      Affairs and Environment
                    Elizabeth Towers
                    St. John's, Newfoundland
                    G. Paulin
                    Directeur,
                    Director de la Recherche
                    Ministere de 1 'Environnement ,
                    2360 Chemin Ste-Foy
                    Quebec G1V 4H2
                    (418) 643-2073

                    P. K. Misra *
                    Air Quality and Meteorology Section
                    Air Resources Branch
                    Ontario Ministry of Environment
                    880 Bay Street,  4th Floor
                    Toronto, Ontario,  M5S1Z8
                    (416) 965-5068

Liaison:            R. Beaulieu
                    United States Transboundary
                      Relations Division
                    Department of External Affairs
                    125 Sussex Drive
                    Ottawa, Ontario, K1AOG2
                    (613) 996-6620

                    Hans Martin
                    LRTAP Liaison Office
                    Atmospheric Environment Service
                    4905 Dufferin Street
                    Downsview, Ontario, M3H5T4
                    (416) 667-4824
  Indicates new member of the Work Group

-------
    Appendix 3
Glossary of Terms

-------
                            A..3-1






Introductory Comments



    During the preparation of this glossary, use has been



made of terminology and definitions found in, inter alia, the



first two annual reports of the United States-Canada Research



Consultation Group on the Long Range Transport of Air Pollutants,



and the draft Federal Acid Rain Assessment Plan.  An obvious



need exists for uniformity in terminology amongst all Work



Groups and others involved in activities related to the



Memorandum of Intent and subsequent developments.  It is



anticipated that this glossary will grow and be refined as



further contributions from specialists in various disciplines



are received.

-------
                           A.3-2






Acid Deposition;  Collectively, the processes by which acidic



and acidifying materials are removed from the atmosphere and



deposited at the surface of the earth.  Also, the amount of



material so deposited. (Units: ML~2T"~1.)



Acid Precipitation;  A more precise term than acid rain, it



usually refers to all types of precipitation with pH less




than 5.6.



Acid Rain; A popular term used to describe precipitation that



is more acidic than "clean" rain (pH   5.6).  It is also used



more generally to describe other atmospheric depositipn



phenomena involving acidity.



Analytical Model;  A mathematical model in which the solution



to the system of governing equations is expressed in terms of



analytical functions.  As such, these models are simplifications



of Lagrangian, Eulerian or statistical models.



Anthropogenic;  Produced by man's activity.



Bulk Deposition;  The term applied to atmospheric deposition



collected in a collector which is open at all times.  Bulk



deposition consists of wet deposition, plus an unknown fraction



of the dry particulate deposition, plus an unknown and probably



very small fraction of the dry gaseous deposition.



Dry Deposition;  Collectively, the processes, excluding preci-



pitation processes, by which materials are removed from the



atmosphere and deposited at the surface of the earth.  Processes



include sedimentation of large particles, the turbulent transfer

-------
                           A.3-3






to the surface of small particles and gases, followed,



respectively, by impaction and sorption or reaction.  Also,



the" amount of material so deposited. (Units: ML~2T~1.)



Ensemble Mean;  The average over a number of individual



model runs in which only one or a few adjustable parameters



are allowed to change.



Eulerian Model; A mathematical model in which computations



are made successively at fixed points in space (as opposed to



Lagrangian models where computations are made following an air



parcel).   Computation points are usually arranged in a fixed



grid, and the model is also known as a grid model.



Flux;  A physical quantity, the amount (mass) of material



passing through ,a unit area in a unit of time.  (Units:



ML-2
-------
                           A.3-4






Isopleth;  A line drawn on a field of values which joins



points of equal value in time or space.



Lagrangian Model;  A mathematical model in which computations



are made successively in the same air parcel(s) as it moves



along a trajectory.  Because this type of model is based on



following an air parcel, it is also known as a trajectory model.



Loading (atmospheric);  The amount of a pollutant in the atmos-



phere expressed in mass or concentration units.  (May also be



expressed on a per unit time and/or area basis.)



Loading Surface;  A term used interchangeably with deposition.



LRTAP;  The long-range transport of air pollutants refers to



the processes, collectively, by which pollutants are transported,



transformed and deposited, on a regional scale (of the order of



hundreds to thousands of km).



Mb (Millibar) Level;  A surface of constant pressure in the



atmosphere, identified by the pressure expressed in mb.



(Common pressure levels used in air quality modeling are 925



and 850 mb levels.)



Mixing Height;  The height above the earth's surface of a



boundary layer inversion which is usually the upper limit of



turbulent mixing activity, and which inhibits upward flux of



pollutant.



Model;  A quantitative simulation of the behaviour of



a portion of the environment.

-------
                           A.3-5





Model Evaluation;  A procedure by which the validity and sen-



sitivity of a model is assessed.  Usually the validity is



ascertained by comparing model outputs with measurements,



and the sensitivity assessed through a series of model runs



in which input parameter values are altered in sequence, and



the results intercompared.



Model Intercomparison;  A procedure of comparing the results



of several models which have been run on specified data bases



and with (usually) specified values of model parameters.



Model Resolution;  The ability of a model to distinguish



(utilize) small spatial or temporal changes in input variables.



Model Sensitivity;  A model characteristic which is described



by the response of an output parameter to a unit change in an



input variable or a model parameter.



Model Validation;  The part of model evaluation in which modeled



results are compared with measured values.



Oxides of Nitrogen;  This term usually denotes the sum of nitric



oxide (NO) and nitrogen dioxide (NC>2).  Other forms are



nitrate (NOj), nitrous oxide (N20), and dinitrogen pentoxide



(N205).



Oxides of Sulfur;  This term usually denotes sulfur dioxide



(S02).  Other forms are sulfur trioxide (303) which is uncommon,



and sulfate (S04).



Parameterization;  The representation of a physical, chemical



or other process by a convenient mathematical expression



containing quantities (parameters) for which measurements or



estimates are usually available.

-------
                          A.3-6





Receptor;  An organism, ecosystem or object which is the



direct or indirect recipient of atmospheric deposition.



Scavenging;  The processes by which materials are incorporated



into precipitation elements and (usually) brought to the earth's



surface.



Scenario;  In the modeling context, a set of specified conditions



(usually emissions inventory) for input to the model which usually



reflect some anticipated future situation (e.g., energy use or



pollution emissions).



Sensitive Area;  A geographical area in which a receptor (or



receptors) exhibit damage in response to a (pollution-imposed)



stress.



Sensitivity Receptor;  The degree to which a receptor exhibits



an adverse effect from a (pollution-imposed) stress.



Source-Receptor Relationship; An expression of how a pollution-



source area and a receptor region are quantitatively linked.



Spatial Resolution;  The minimum distance in space over which



meaningful differences in results can be determined (using a



particular model.)   (For example, a model based on a 381-km



grid will provide no significantly different information for



two receptor points  separated by less than approximately 381 km.)



Stationarity;  Turbulent flow field is stationary when the flow



characteristics remain independent of the initial conditions.



Statistical Model; A mathematical model which uses statistical



values of parameters as inputs for the computations.

-------
                           A.3-7





Surrogate;  The term applied to a parameter which is used to



represent another.  (For example, modeling hydrogen ion



behavior in the atmosphere is difficult, so that sulfate ion



is used as a substitute.)



Susceptibility:  A receptor or receptor area is said to be



susceptible if it is both sensitive, and receiving a pollutant



loading or stress.



Temporal Resolution;  The minimum time during which meaningful



differences in results can be determined (using a particular



model).  (For example, models using upper air data which are



only available every six hours are limited in their temporal



resolution to.about 6 hours.)



Trajectory;  The path or track of an air parcel through the



atmosphere.  It can be calculated from observed or gridded



wind data either forward or backward from a point (source or



receptor, respectively).



Transfer Matrix;  A presentation of source-receptor relation-



ships in a matrix form.  Matrix elements can be expressed



as percentage values, as absolute values, or as values



normalized by source strength.)  Such a presentation provides



a means of easy comparison of the impact of a variety of



sources on a variety of receptors.



Transformation (chemical);  The processes by which chemical



species are converted into other chemical species (in the



atmosphere).

-------
                           A.3-8






Variance;   A measure of variability.  It is denoted by o 2




and defined as the mean-square deviation from the mean, that



is, the mean of the squares of the differences between



individual values of x and the mean value 3c.



     6"^ = E [(x-x)2], where E denotes the expected value.




Wet Deposition;  Collectively, the processes by which materials



are removed from the atmosphere and deposited at the surface



of the earth by precipitation elements.  The processes include



in-cloud and below-cloud scavenging of both gaseous and



particulate materials.  Also, the amount of material so



deposited. (Units: ML~2T~1.)

-------
                  APPENDIX 4
Compilation of Attendance, Agenda and Minutes
   for the Modeling Subgroup Workshops and
            Work Group 2 Meetings

-------
                            A.4-1


Document 2-16 Available From:
James W.S. Young
Air Quality and Inter-Environmental
  Research Branch
Atmospheric Environment Service
Environment Canada
4905 Dufferin Street
Downsview, Ontario M3H5T4
(416) 667-4786
Brand L. Niemann
Program Integration and Policy Staff (RD-681)
Office of Research and Development
Environmental Protection Agency
Washington, D.C.  20460
(202) 426-9434

-------
         APPENDIX 5

List of Work Group 2 Reports
    and Other Documents

-------
                               A. 5-1
          List of Work Group 2 Reports and Other Documents

                              Revised
                           (July 3, 1981)
                                                         Identification
Phase
I


Document Name
WG2 Phase I Interim Report
Addendum to Appendix 6
Addendum to Appendix 8
Date
1/14/81
1/14/81
2/20/81
Number
2-1
2-2
2-3
II     Unified S02 Emission Inventory
       (1976-1980)
       AES-LRT Model Profile
       ASTRAP Model Profile
       ENAMAP Model Profile
       OME-LRT Model Profile
       RCDM Model Profile
       UMACID Model Profile
       MEP-TRANS Model Profile
       CAPITA-Monte Carlo Model Profile
6/30/81

5/15/81
5/12/81
6/30/81
3/31/81
7/10/81
6/24/81 (Interim)
6/30/81 (Interim)
6/30/81 (Interim)
       Modeling Sub-Group Phase II
         Report                            7/10/81
       Atmospheric Sciences Sub-Group
         Phas<% II Report                   7/10/81
           Sulfur and Nitrogen Chemistry
             in LRT Models
           Trends in Precipitation
             Composition and Deposition
           Seasonal Dependence of Atmos-
             pheric Deposition and Chemical
             Transformation Rates for Sulfur
             and Nitrogen Compounds
           Global Distribution of Acidic
             Precipitation and the
             Implications for Eastern
             North America
       WG2 Phase II "Working Report"

       Compilation of Attendance,
             Agenda, and Minutes
7/10/81


7/10/81
2-4

2-5
2-6
2-7
2-8
2-9
2-10
2-11
2-12
2-13

2-14

2-14A

2-14B



2-14C



2-14D

2-15


2-16

-------
                    APPENDIX 6

      A Simple Example of the Role of Models
in the Development of Emission Control Strategies

-------
                            A.6-1

A SIMPLE EXAMPLE OF THE ROLE OF MODELS IN THE DEVELOPMENT
OF EMISSION CONTROL STRATEGIES

A6.1  Introduction
     This section outlines the use of transfer matrices

generated by the models for evaluating emission limits for

selected source regons given the desired air quality objectives.

Transfer matrices are based on the assumption that the average

concentration or deposition of a pollutant in any receptor

area is a linear combination of emissions of its precursor in

every source region.  For example, deposition at receptor 1

is given by
                + T12E2 + Tl3E3  + ............ + T].NEN   (A6.1)

where Ej is the emission rate in region j and T]_j are a set

of transfer coefficients.  A similar relationship for deposition

holds for other receptors except that a different set of

coefficients T^j will apply.  Thus a two-dimensional array of

transfer coefficients (a matrix) is required to describe the

deposition values at all receptors in terms of emissions from

all source regions.

     Using matrix notation the relationship between sources

and deposition may be expressed as

     If . T = D
             S\s

or                                                      (A6.2)

     N
    ^"E-jTi-; = Dif i = 1, ...... M
     3=1

where N is the total number of source regions and M the total
number of receptors.

-------
                            A..6-2



     The transfer matrix T is generated using a model/ or


models, of long-range atmospheric transport to simulate the


movement of pollutants emitted from sources in each emitter

             •9
region as they are dispersed across the receptor areas.  Then


the matrix element T^j is calculated by separating out the


deposition produced at receptor i by emission in region j and


dividing by the magnitude of that emission.  The transfer


matrix may then be combined with selected emission vectors E


representing various emission scenarios to assess the impact


on the deposition array D.


A6. 2  Illustration of Transfer Matrix Use in Assessment


     To show the use of transfer matrices, a simple two source


and two receptor example is now presented.  The problem to be


solved is - what are the allowable emissions for the source


regions which permit attainment to deposition limits at the


receptors and which meet the constraints on the magnitude of


the emissions?


     The target depositions and the constraints on the


emissions for the two dimensional problem may be expressed


mathematically as


                                                       (A6.3)
Ei\  ^ / EMAXi
                                                       (A6.4)
where EMINj is the minimum emission region j is able to attain

-------
                            A.6-3


and EMAXj is the maximum emission region j is permitted.  These

constraints define a set of feasible solutions.  Graphically

these feasible solutions form an area in the two dimensional

variable space of (Ej_, £2).  The boundaries of the areas are

defined by the six lines in (Ej_, £2) space defined in Equations

(A6.3) and (A6.4), namely:

     E1T11  +  E2T12 = Dl         (line 1)

     E1T12  +  E2T22 = D2         (line 2)

                                                         (A5.5)
     EI = EMIt^                   (line 3)

     £2 = EMIN2                   (line 4)

     EI = EMAX;L                   (line 5)

     E2 = EMAX2                   (line 6)
                                                           »

Figure A6-1 shows a schematic diagram of the region of feasible

solutions as defined by a set of lines such as those given in

lines 3 to 5.

     Although the sequence of equations in lines 3 to 5

define the set of feasible solutions the selection of an

optimal solution requires a further criterion.  This criterion

will define the objective of the controls strategy which is

usually the minimization of abatement costs while meeting

the constraints on deposition and emissions.  For example,

assume that control costs are directly proportional to the

magnitude of the reduction in emission.  Hence, to minimize

cost, the emissions must be maximized, i.e.

     •Maximize
     subject to
     constraint
= G(E)    (line 7)    (A6.6)

-------
                                                A. 6-4
       \
        \
          \
            \
                  \
                       (line 3)
                         (line 2)
                           E]T2l
             E2T22 =
                             (line 5)
                          \
                             optimal feasible
                              solution (E]_*, E2*)
EMAX2
EMIN2
                                                     infinite
                                                       t of
                                                     parallel  lines
                                                     defined by
    FEASIBLE^SOLUT ION
                                                                                 (line 6)
                                                                                  E2  = EMAX2
(line 1)
     E1T11+E2T12=D1
                                                                            S j E = csDnstant
                                                          (line 4)
                                                          E2 = EMIN2
                                                 EMAXi
        Figure A6-1:
Graphic representation of control strategy assessment,
Lines 1-6 define the two-dimensional solution space
of the control strategy problem.  The diagonal solid
lines denote solutions to the objective function for
the control scenario and the dot marks the optimal
solution (E]_*, E2*) for the example.

-------
                            A.6-5






     Equation (A6.6) is the objective function which is used



to select the optimal solution.  The relationship




     N,

    "^_I E-; = constant


    j=l
                                      t


defines an infinite set of parallel lines through the region



of feasible solutions for which G(E) is the constant describing



the line intersecting the region at the optimal solution  (or



solution).  The values of Ej_ and £2 at the point marked by



the dot in Figure A6-1 are solution to this control strategy



exmaple.  It is evident from this two-dimensional example



that the line passing through the dot has the maximum intercepts



on EI and £2 axes, of all lines parallel to it and having a



segment in the shaded region.  This establishes the optimality



of .the feasible solution.

-------